8.4.1 Server environment definition (servers)
You create a definition file containing the server environment definitions under the file name servers in the directory for HA Monitor environment settings.
A sample servers file is provided under the directory for HA Monitor sample files. If you copy this file to the directory for HA Monitor environment settings and then edit it, you can save time over creating the file from scratch.
You must create this definition file for each host. Bold type indicates a value that must be the same on all hosts. Care must be taken in specifying the other items so that there will be no conflicts between the hosts.
The following shows the format of the definition file used to specify server environment definitions:
/* Server environment definition */ server name program-name ,alias server-alias-name [,initial {online|standby}] [,standbypri {priority-of-each-standby-server|1}] [,group group-name[:{exchange|no_exchange[: {cancel|no_cancel}]}]] [,parent alias-name-of-parent-server] [,acttype {server|monitor}] [,program {use|nouse}] [,patrol server-failure-monitoring-time] [,ph_threshold time-when-server-monitoring-history-is-to-be-obtained] [,actcommand absolute-path-name-of-server-start-command] [,termcommand absolute-path-name-of-server-termination-command] [,patrolcommand absolute-path-name-of-server-monitoring-command] [,ptrlcmd_ex absolute-path-name-of-server-monitoring-command] [,ptrlcmd_ex_inter server-monitoring-command-execution-interval] [,ptrlcmd_ex_retry {server-monitoring-command-retry-count|2}] [,ptrlcmd_ex_tmout server-monitoring-command-timeout-value] [,start_timeout timeout-for-server-startup] [,stop_timeout timeout-at-server-termination] [,stop_timeout_sw timeout-at-server-termination] [,waitserv_exec {yes|no}] [,servexec_retry {server-startup-processing-retry-count|2}] [,retry_stable {startup-completion-monitoring-time|60}] [,sby_actcommand path-name-of-standby-server-start-command] [,sby_termcommand path-name-of-standby-server-termination-command] [,sby_ptrlcmd_ex absolute-path-name-of-standby-server-monitoring-command] [,sby_ptrlcmd_ex_inter execution-interval-of-standby-server-monitoring-command [,sby_ptrlcmd_ex_retry retry-count-of-standby-server-monitoring-command|2] [,sby_ptrlcmd_ex_tmout timeout-value-for-execution-of-standby-server-monitoring-command] [,sby_start_timeout timeout-value-for-processing-of-standby-server-start-command] [,sby_stop_timeout timeout-value-for-processing-of-standby-server-termination-command] [,sby_servexec_retry retry-count-for-standby-server-startup-processing|0] [,switchtype {switch|restart[:restart-monitoring-time]|manual}] [,switch_error {retry|ignore}] [,switch_retry {start-retry-count|2}] [,switchbyfail LAN-interface-name[:LAN-interface-name...]] [,disk_ptrl {use|nouse}] [,disk_ptrl_inter {check-interval-for-disk-for-business-use|120}] [,disk_ptrl_retry {check-retry-interval-for-disk-for-business-use|60}: {check-retry-count-for-disk-for-business-use|1}] [,disk_ptrl_act {switch|pairdown|noswitch}] [,switch_judge LAN-interface-name] [,pairdown {use[:serv_slow]|nouse}] [,uoc_neck {use|nouse}] [,uoc_pairdown {use|nouse}] [,term_pairdown {use|nouse}] [,lan_updown {use|nouse}] [,ip_neck {use|nouse}] [,disk absolute-path-name-of-volume-group[:absolute-path-name-of-volume-group...]] [,vg_on_opt volume-group-connection-option[:option...]] [,vg_neck {use|nouse}[: {use|nouse}...]] [,fs_name absolute-path-name-of-logical-volume[:absolute-path-name-of-logical-volume...]] [,fs_mount_dir directory-name[:directory-name...]] [,fs_mount_opt option[:option...]] [,fs_umount_retry {retry-count|10}] [,fs_neck {use|nouse}[: {use|nouse}...]] [,scsi_device device-name[:device-name...]] [,dmmp_device device-name[:device-name...]] [,deviceoff_order {order|reverse}] [,dev_timelimit timeout-value-for-switchover-of-shared-resources] [,dev_offlimit timeout-value-for-disconnection-of-resources-shared-by-servers] [,dev_onlimit timeout-value-for-connection-of-resources-shared-by-servers]; /* Resource server environment definition */ [resource alias alias-name-of-resource-server [,initial {online|standby}] [,standbypri {priority-of-each-standby-server|1}] ,group group-name [,switch_error {retry|ignore}] [,switch_retry {start-retry-count|2}] [,switchbyfail LAN-interface-name[:LAN-interface-name...]] [,switch_judge LAN-interface-name] [,uoc_neck {use|nouse}] [,uoc_pairdown {use|nouse}] [,lan_updown {use|nouse}] [,ip_neck {use|nouse}] [,disk absolute-path-name-for-volume-group[:absolute-path-name-for-volume-group...]] [,vg_on_opt volume-group-connection-option[:option...]] [,vg_neck {use|nouse}[: {use|nouse}...]] [,fs_name path-name-for-file-system[:path-name-for-file-system...]] [,fs_mount_dir directory-name[:directory-name...]] [,fs_mount_opt option[:option...]] [,fs_umount_retry {retry-count|10}] [,fs_neck {use|nouse}[: {use|nouse}...]] [,scsi_device device-name[:device-name...]] [,dmmp_device device-name[:device-name...]] [,deviceoff_order {order|reverse}] [,dev_timelimit timeout-value-for-switchover-of-shared-resources] [,dev_offlimit timeout-value-for-disconnection-of-resources-shared-by-servers] [,dev_onlimit timeout-value-for-connection-of-resources-shared-by-servers] [,disk_ptrl {use|nouse}] [,disk_ptrl_inter {check-interval-for-disk-for-business-use|120}] [,disk_ptrl_retry {check-retry-interval-for-disk-for-business-use|60}: {check-retry-count-for-disk-for-business-use|1}] [,disk_ptrl_act {switch|pairdown|noswitch}];
Do not specify any of the following operands when the function for controlling hot standby based on the availability of LAN communications is used by business applications (use is specified in the fence_lan operand):
-
disk
-
vg_neck
-
fs_name
-
fs_mount_dir
-
fs_mount_opt
-
fs_umount_retry
-
fs_neck
-
scsi_device
-
dmmp_device
- Organization of this subsection
(1) List of server types and operands to be defined
The operands to be defined in a server environment definition depend on the server type. The server types are listed below:
-
Server in the server mode
-
Server in the monitor mode
-
Resource server
The following table shows the applicable operands for each server type.
Operand |
Server mode |
Monitor mode |
Resource server |
|
---|---|---|---|---|
For TP1/Server Base or HiRDB |
For TP1/EE |
|||
name |
R |
R |
-- |
|
alias |
R |
R |
R |
|
initial |
D |
D |
D |
|
standbypri |
D |
D |
D |
|
group |
O |
O |
R |
|
parent |
O |
O |
-- |
|
acttype |
D |
R |
-- |
|
program |
-- |
D |
-- |
|
patrol |
R |
-- |
-- |
-- |
ph_threshold |
O |
-- |
-- |
-- |
actcommand |
-- |
O |
-- |
|
termcommand |
-- |
O |
-- |
|
patrolcommand |
-- |
O |
-- |
|
ptrlcmd_ex |
-- |
O |
-- |
|
ptrlcmd_ex_inter |
-- |
O#2 |
-- |
|
ptrlcmd_ex_retry |
-- |
D |
-- |
|
ptrlcmd_ex_tmout |
-- |
O |
-- |
|
start_timeout |
-- |
O |
-- |
|
stop_timeout |
-- |
O |
-- |
|
stop_timeout_sw |
-- |
O |
-- |
|
waitserv_exec |
-- |
D |
-- |
|
servexec_retry |
-- |
D |
-- |
|
retry_stable |
-- |
D#3 |
-- |
|
sby_actcommand |
-- |
O |
-- |
|
sby_termcommand |
-- |
O |
-- |
|
sby_ptrlcmd_ex |
-- |
O |
-- |
|
sby_ptrlcmd_ex_inter |
-- |
O#4 |
-- |
|
sby_ptrlcmd_ex_retry |
-- |
D |
-- |
|
sby_ptrlcmd_ex_tmout |
-- |
O |
-- |
|
sby_start_timeout |
-- |
O |
-- |
|
sby_stop_timeout |
-- |
O |
-- |
|
sby_servexec_retry |
-- |
D |
-- |
|
switchtype |
D |
-- |
-- |
|
switch_error |
D |
D |
D |
|
switch_retry |
-- |
D |
D |
|
switchbyfail |
O |
O |
O |
|
disk_ptrl |
D |
D |
D |
|
disk_ptrl_inter |
D |
D |
D |
|
disk_ptrl_retry |
D |
D |
D |
|
disk_ptrl_act |
D |
D |
D |
|
switch_judge |
O |
O |
O |
|
pairdown |
D |
--#1 |
-- |
|
uoc_neck |
D |
D |
D |
|
uoc_pairdown |
D |
D |
D |
|
term_pairdown |
-- |
O |
-- |
|
lan_updown |
D |
D |
D |
|
ip_neck |
D |
D |
D |
|
disk |
O |
O |
O |
|
vg_on_opt |
D |
D |
D |
|
vg_neck |
D |
D |
D |
|
fs_name |
O |
O |
O |
|
fs_mount_dir |
O |
O |
O |
|
fs_mount_opt |
O |
O |
O |
|
fs_umount_retry |
D |
D |
D |
|
fs_neck |
D |
D |
D |
|
scsi_device |
O |
O |
O |
|
dmmp_device |
O |
O |
O |
|
deviceoff_order |
D |
D |
D |
|
dev_timelimit |
O |
O |
O |
|
dev_offlimit |
O |
O |
O |
|
dev_onlimit |
O |
O |
O |
- Legend:
-
R: Required (cannot be omitted).
D: Default value is assumed if specification is omitted.
O: Optional (no value is assumed if left unspecified).
--: Not applicable (ignored, if defined).
- #1
-
O applies if ptrlcmd_ex is specified.
- #2
-
R applies if ptrlcmd_ex is specified.
- #3
-
-- applies if ptrlcmd_ex is specified.
- #4
-
R applies if sby_ptrlcmd_ex is specified.
(2) server definition statement
The server definition statement defines a server's operating environment. The following explains the operands of the server definition statement.
(a) name
User-specified values
Value type |
Range of numerical values |
Can be omitted |
Assumed value when this value is omitted |
---|---|---|---|
Server path name having 1 to 1000 characters |
-- (Units: --) |
No |
-- |
Description
Specifies a name used to identify the program.
Specify the same value for all hosts in the hot-standby configuration.
- For a server in the server mode:
-
-
When using TP1/Server Base
Specify the absolute path name of the OpenTP1 directory. For details about the OpenTP1 directory, see the manual OpenTP1 Operation.
-
When using HiRDB
The value depends on the HiRDB system switchover facility that is being used. For details about the value to be specified, see the HiRDB System Operation Guide (UNIX).
-
- For a server in the monitor mode:
-
Specify the absolute path name of the server start command. If the actcommand operand is specified, specify either the absolute path name of server start command or a unique name that can identify the program.
Notes
Do not specify an asterisk (*) at the end of the operand when using the online serverless mode.
(b) alias
User-specified values
Value type |
Range of numerical values |
Can be omitted |
Assumed value when this value is omitted |
---|---|---|---|
1 to 8 alphanumeric characters |
-- (Units: --) |
No |
-- |
Description
Specifies the server alias name that is to be used in HA Monitor commands and in the messages that are issued.
This name must be unique within each host. Specify the same value among all hosts in a hot-standby configuration.
(c) initial
User-specified values
Value type |
Range of numerical values |
Can be omitted |
Assumed value when this value is omitted |
---|---|---|---|
online|standby |
-- (Units: --) |
Yes |
online |
Description
Specifies the server's start status, which establishes the server's type at the time of startup.
Specify online for the primary system and standby for the secondary system.
-
online: Starts the server as the active server.
-
standby: Starts the server as a standby server.
(d) standbypri
User-specified values
Value type |
Range of numerical values |
Can be omitted |
Assumed value when this value is omitted |
---|---|---|---|
Unsigned integer |
1 to 9999 (Units: --) |
Yes |
1 |
Description
This operand specifies the priority of each standby server if the multi-standby function is used.
The smaller the value of this operand, the higher the priority of the host. You must specify a different value for each host. This operand becomes effective only when use is specified in the multistandby operand in the HA Monitor environment settings and standby is specified in the initial operand in the server environment definition. If online is specified in the initial operand, the highest priority is assumed. For details about the multi-standby function, see 4.5 Managing servers and hosts when using the multi-standby function.
(e) group
User-specified values
Value type |
Range of numerical values |
Can be omitted |
Assumed value when this value is omitted |
---|---|---|---|
Server group name: 1 to 8 alphanumeric characters |
-- (Units: --) |
Yes |
-- |
Server switchover type: exchange|no_exchange |
-- (Units: --) |
Yes |
no_exchange# |
#: For monitor-mode servers, no default value is provided.
Description
Specifies the name of the server group to be switched over in the batch mode during grouped-system switchover and the server's switchover type.
Specify the same value for all hosts in the hot-standby configuration. When you specify the group operand, you must specify the same initial operand value for all grouped servers on the same host.
You cannot specify exchange, no_exchange, cancel, or no_cancel for a server group name.
For the server switchover type, specify whether a grouped-system switchover is to be performed on the server in the event of a server failure. You can specify the server switchover type for servers in the server mode.
The server switchover type is valid only for servers in the server mode. The default server switchover type is no_exchange. Specify exchange if you want grouped-system switchover to be performed in the event of a server failure.
The server switchover type cannot be specified for a server in the monitor mode. In the event of a server failure on a server in the monitor mode, if a server monitoring command has been created, grouped-system switchover will be performed. If no server monitoring command has been created, grouped-system switchover will not be performed.
Specification of a server switchover type is optional. If you specify a server switchover type, you must also specify a group name. If you omit the group name, the specified switchover type for the server will be ignored. You can specify either of the following values for the server switchover type:
-
exchange: If a server failure occurs on the server (when this value is specified), grouped-system switchover is performed on all the servers in the same server group in the batch mode. This value and cancel or no_cancel are mutually exclusive.
-
no_exchange: If a server failure occurs on the server (when this value is specified), grouped-system switchover is not performed. The corresponding standby server is placed in the grouped-system switchover wait state. However, grouped-system switchover is performed in the following cases:
-
no_exchange is specified for all servers of the server group, and a server failure occurs on all the servers.
-
The server group includes a resource server and all the resource server's child servers fail.
If TP1/Server Base or HiRDB is used as the server, you can specify one of the values shown later after no_exchange. If you specify both no_exchange and cancel or no_cancel, separate the two values with a colon (:).
-
- cancel
-
If HA Monitor detects a slowdown on the server when this value is specified, HA Monitor terminates the server abnormally. The corresponding standby server is placed in the grouped-system switchover wait state. However, if manual is specified for the switchtype operand, HA Monitor continues to monitor the status of this server.
- no_cancel
-
If HA Monitor detects slowdown on the server when this value is specified, HA Monitor continues monitoring the server. The corresponding standby server is not placed in the grouped-system switchover wait state.
If no_exchange:no_cancel is specified for all servers in the server group and HA Monitor detects a slowdown on all servers, HA Monitor continues monitoring server status without performing hot standby processing.
(f) parent
User-specified values
Value type |
Range of numerical values |
Can be omitted |
Assumed value when this value is omitted |
---|---|---|---|
1 to 8 alphanumeric characters |
-- (Units: --) |
Yes |
-- |
Description
Specifies the parent server required for starting the local server when the server switchover order is to be controlled.
Specify the same value for all hosts in the hot-standby configuration. In this operand, specify the server alias name (alias) of the parent server. When this operand is specified, HA Monitor controls the start order in such a manner that a local server is not started until its parent server's startup processing has been completed. You can specify this operand only if the group operand is specified.
Do not change this operand's value while a server in the server group is running. If it is changed, HA Monitor can no longer control the start order correctly.
(g) acttype
User-specified values
Value type |
Range of numerical values |
Can be omitted |
Assumed value when this value is omitted |
---|---|---|---|
server|monitor |
-- (Units: --) |
Yes |
server |
Description
Specifies the server's start method.
For a server in the server mode, specify server; for a server in the monitor mode, specify monitor. Specify the same value for all hosts in the hot-standby configuration.
-
server: Starts the server by using the server start command.
-
monitor: Starts the server by using the HA Monitor command.
(h) program
User-specified values
Value type |
Range of numerical values |
Can be omitted |
Assumed value when this value is omitted |
---|---|---|---|
use|nouse |
-- (Units: --) |
Yes |
nouse |
Description
Specifies whether the monitor-mode program management function is to be used.
You must specify the same value for all hosts in the hot-standby configuration.
-
use: Uses the monitor-mode program management function.
-
nouse: Does not use the monitor-mode program management function.
If you specify use in this operand, also specify the following definitions:
-
pgmmax operand in the HA Monitor environment settings
-
Monitor-mode program environment definition
For details about the monitor-mode program environment definition, see 8.5.1 Monitor-mode program environment definition (programs).
For a server in the server mode, this operand is ignored, if specified.
(i) patrol
User-specified values
Value type |
Range of numerical values |
Can be omitted |
Assumed value when this value is omitted |
---|---|---|---|
Unsigned integer |
5 to 600 (Units: seconds) |
Yes |
-- |
Description
Specifies the interval (server failure monitoring time) at which HA Monitor is to check for an operation report from the server (an operation report provides criteria for detecting server failures).
You must specify the same value for all hosts in the hot-standby switchover configuration. If no operation report is received from the server within the amount of time specified in the patrol operand, HA Monitor assumes that a server failure occurred.
If TP1/EE is used as a server-mode server or if a monitor-mode server is used, specification of this operand is ignored.
(j) ph_threshold
User-specified values
Value type |
Range of numerical values |
Can be omitted |
Assumed value when this value is omitted |
---|---|---|---|
Unsigned integer |
5 to 559 (Units: seconds) |
Yes |
--# |
#: If this operand is omitted, server monitoring history is not collected.
Description
When information about slowdowns of the local host is to be collected as monitoring history, this operand specifies the amount of time before collection of monitoring history is to begin after the active server stops sending operation reports (time when server monitoring history is to be obtained).
You must specify the same value for all hosts in the hot-standby configuration.
The value for the time when server monitoring history is to be obtained must be less than the value for the server failure monitoring time specified in the patrol operand in the server environment definition.
If a monitor-mode server is used, specification of this operand is ignored.
When you specify this operand to collect server monitoring history, also specify the ph_log_size operand.
(k) actcommand
User-specified values
Value type |
Range of numerical values |
Can be omitted |
Assumed value when this value is omitted |
---|---|---|---|
Path name having 1 to 1000 characters |
-- (Units: --) |
Yes |
-- |
Description
Specifies the absolute path name of the server's start command.
As with the termcommand operand, you can specify arguments by enclosing the operand value in double quotation marks ("). A double quotation mark is not counted as a character. Specify the same value for all hosts in the hot-standby configuration.
For a server in the monitor mode, if you want to specify arguments in the server start command, specify this operand. If you specify this operand, there is no need to specify the absolute path name of the server start command in the name operand. In the name operand, specify a unique name that can identify the program
For a server in the server mode, this operand is ignored, if specified.
(l) termcommand
User-specified values
Value type |
Range of numerical values |
Can be omitted |
Assumed value when this value is omitted |
---|---|---|---|
Path name having 1 to 1000 characters |
-- (Units: --) |
Yes |
-- |
Description
Specifies the absolute path name of a server termination command that has been created by using a shell script, for example.
Specify the same value for all hosts in the hot-standby configuration.
If you specify this operand for a server in the monitor mode, you can have planned termination performed automatically on the active server at any time simply by executing the server hot-standby switchover command (monswap command). You can also have the active server terminated normally at any time simply by executing the monitor-mode server termination command (monend command).
For a server in the server mode, this operand is ignored, if specified.
You can include spaces in the absolute path name by enclosing the entire name in double quotation marks ("). By using spaces, you can specify arguments for the command. A double quotation mark (") is not counted as a character. If you use a specification that includes spaces, the first character string before the first space is recognized as the path name of the server termination command. The character strings after the first space are recognized as arguments. An argument passed by HA Monitor (-e, -w, -c) is added and executed as the last argument. For details about the arguments that are passed by HA Monitor, see 6.13.2 Creating a server termination command.
(m) patrolcommand
User-specified values
Value type |
Range of numerical values |
Can be omitted |
Assumed value when this value is omitted |
---|---|---|---|
Path name having 1 to 1000 characters |
-- (Units: --) |
Yes |
--# |
- #
-
If this operand is omitted, HA Monitor does not monitor the active server.
However, if the ptrlcmd_ex operand is specified, HA Monitor monitors the active server.
Description
Specifies the absolute path name of a server monitoring command to be used to monitor the server in the monitor mode.
Specify the same value for all hosts in the hot-standby configuration.
For a server in the server mode, this operand is ignored, if specified.
As with the termcommand operand, you can specify arguments in this operand by enclosing the operand value in double quotation marks (").
(n) ptrlcmd_ex
User-specified values
Value type |
Range of numerical values |
Can be omitted |
Assumed value when this value is omitted |
---|---|---|---|
Path name having 1 to 1000 characters |
-- (Units: --) |
Yes |
--# |
- #
-
If this operand is omitted, HA Monitor does not monitor the active server.
However, if the patrolcommand operand is specified, HA Monitor monitors the active server.
Description
Specifies the absolute path name of a server monitoring command to be used to monitor the server in the monitor mode.
Specify the same value for all hosts in the hot-standby configuration.
For a server in the server mode, this operand is ignored, if specified.
Note that you can include arguments in the path name by enclosing it in double quotation marks (").
If you specify this operand, you must also specify the ptrlcmd_ex_inter operand. If necessary, specify the ptrlcmd_ex_tmout operand. Note that if you specify this operand, you cannot specify the patrolcommand operand. Also, you cannot specify the program operand.
(o) ptrlcmd_ex_inter
User-specified values
Value type |
Range of numerical values |
Can be omitted |
Assumed value when this value is omitted |
---|---|---|---|
Unsigned integer |
3 to 180 (Units: seconds) |
Yes# |
-- |
- #
-
You must specify this operand if you specify the ptrlcmd_ex operand.
Description
Specifies the interval at which the server monitoring command is executed to monitor the server in the monitor mode. Make sure that the interval you specify is longer than the time period from the start to end of the server monitoring command.
You must specify the same value for all hosts in the hot-standby switchover configuration.
For a server in the server mode, this operand is ignored, if specified.
(p) ptrlcmd_ex_retry
User-specified values
Value type |
Range of numerical values |
Can be omitted |
Assumed value when this value is omitted |
---|---|---|---|
Unsigned integer |
0 to 19 (Units: number of times) |
Yes |
2 |
Description
If the server monitoring command executed to monitor a monitor-mode server fails (with the EXIT code 1 to 9), command execution is retried by the number of times specified for this operand.
You must specify the same value for all hosts in the hot-standby switchover configuration.
If an attempt to execute the command fails the specified number of times in succession, HA Monitor assumes that the active server failed and operates as specified by the servexec_retry operand.
(q) ptrlcmd_ex_tmout
User-specified values
Value type |
Range of numerical values |
Can be omitted |
Assumed value when this value is omitted |
---|---|---|---|
Unsigned integer |
5 to 3600 (Units: seconds) |
Yes |
-- |
Description
Specifies the timeout value for the server monitoring command that is executed to monitor the server in the monitor mode. Make sure that the timeout value you specify is equal to or longer than the time period from the start to end of the server monitoring command. If you specify a value shorter than the time period from the start to end of the monitoring command, monitoring cannot be performed appropriately because HA Monitor might perform a false detection. If the command times out, HA Monitor assumes that a server failure occurred, and operates as specified by the servexec_retry operand.
You must specify the same value for all hosts in the hot-standby switchover configuration.
(r) start_timeout
User-specified values
Value type |
Range of numerical values |
Can be omitted |
Assumed value when this value is omitted |
---|---|---|---|
Unsigned integer |
60 to 3600 (Units: seconds) |
Yes |
--# |
#: If this operand is omitted, HA Monitor does not monitor the timeout for startup of the server in the monitor mode.
Description
Specifies a timeout value for the server's startup processing in the monitor mode.
We recommend that you specify this operand to prevent indefinite waiting for server startup.
Specify the same value for all hosts in the hot-standby configuration. When this operand is specified, the target server's startup processing is also subject to this timeout monitoring.
If the server is to be restarted according to the specification of the switch_retry operand, the processing is also subject to this timeout monitoring.
If a server's startup processing is retried according to the specification of the servexec_retry operand, its startup processing is not subject to timeout monitoring.
If you specify too small a timeout value, hot-standby switchover might fail to start the other server. To avoid this, specify a sufficiently large value such as a few minutes or the maximum value (3600).
If a timeout occurs, HA Monitor forcibly terminates the server start command and cancels the server's startup processing. If you want to execute a server termination command after a timeout has occurred, specify use in the termcmd_at_abort operand in the HA Monitor environment settings.
If you have specified yes in the waitserv_exec operand, specify the amount of time from start to termination of the server start command. If you have specified use in the program operand, include the amount of time before the hamon_patrolstart function is issued as the timeout value. HA Monitor monitors only once the amount of time before the hamon_patrolstart function is issued. Even when a UAP consists of multiple processes, HA Monitor monitors only once the amount of time before the hamon_patrolstart function is issued.
This operand is ignored if specified for a server in the server mode or when no is specified in the waitserv_exec operand and nouse is specified in the program operand.
The following table shows the correspondence between the waitserv_exec and program operand values and the monitoring range to be specified.
Operand value |
Monitoring range to be specified |
|
---|---|---|
waitserv_exec |
program |
|
yes |
use |
From the start to termination of a server start command and the amount of time before the hamon_patrolstart function is issued |
nouse |
From the start to termination of a server start command |
|
no |
use |
Length of time before the hamon_patrolstart function is issued |
nouse |
Timeout is not monitored (the operand is ignored, if specified). |
(s) stop_timeout
User-specified values
Value type |
Range of numerical values |
Can be omitted |
Assumed value when this value is omitted |
---|---|---|---|
Unsigned integer |
60 to 3600 (Units: seconds) |
Yes |
--# |
#: If this operand is omitted, HA Monitor does not monitor the timeout for termination of the server in the monitor mode.
Description
Specifies a timeout value for the server's termination processing in the monitor mode.
We recommend that you specify this operand to prevent indefinite waiting for server termination.
Specify the same value for all hosts in the hot-standby configuration.
If the server is to be restarted according to the specification of the switch_retry operand, the processing is also subject to this timeout monitoring.
If the server's startup processing is retried according to the specification of the servexec_retry operand, the processing is not subject to timeout monitoring.
For the timeout value, specify the maximum wait time that is allowable for server termination. Make sure that the time you specify is equal to or longer than the time period from the start to end of the server termination command. The value specified here takes effect for normal server termination and for server termination at hot-standby switchover.
If you want to set a smaller timeout value for server termination at hot-standby switchover than for normal server termination, specify the timeout value for server termination at hot-standby switchover for the stop_timeout_sw operand. If you specify both the stop_timeout operand and the stop_timeout_sw operand, the specification of the stop_timeout operand is applied to normal termination, and the stop_timeout_sw operand is applied to termination at hot-standby switchover.
If you have specified use in the program operand, include in the timeout value the amount of time before the hamon_patrolstop function is issued. HA Monitor monitors for the amount of time before the hamon_patrolstop function is issued from all programs that have issued the hamon_patrolstart function.
If a timeout occurs, HA Monitor forcibly terminates the server termination command. If a UAP process is running, HA Monitor also terminates the UAP process.
For a server in the server mode, this operand is ignored, if specified.
(t) stop_timeout_sw
User-specified values
Value type |
Range of numerical values |
Can be omitted |
Assumed value when this value is omitted |
---|---|---|---|
Unsigned integer |
1 to 3600 (Units: seconds) |
Yes |
-- |
Description
Specifies the timeout value for termination of the monitor-mode server at hot-standby switchover.
You must specify the same value for all hosts in the hot-standby switchover configuration.
For the timeout value, specify the maximum wait time that is allowable for server termination at hot-standby switchover. Make sure that the time you specify is equal to or longer than the time period from the start to end of the server termination command.
If you want to set for server termination at hot-standby switchover a smaller timeout value than for normal server termination, specify this operand. If you specify both the stop_timeout operand and the stop_timeout_sw operand, the specification of the stop_timeout operand is applied to normal termination, and the stop_timeout_sw operand is applied to termination at hot-standby switchover.
(u) waitserv_exec
User-specified values
Value type |
Range of numerical values |
Can be omitted |
Assumed value when this value is omitted |
---|---|---|---|
yes|no |
-- (Units: --) |
Yes |
no |
Description
Specifies whether HA Monitor is to wait until execution of the server start command specified in the name operand has been completed before it performs active server start completion processing when a server in the monitor mode is to be started as the active server.
Specify the same value for all hosts in the hot-standby configuration.
-
yes: Waits until execution of the server start command is completed and then performs active server start completion processing.
-
no: Performs the active server start completion processing without waiting for completion of the server start command.
If you are monitoring a server in the monitor mode, determine the value of this operand based on the information provided in (2) How to create a server start command.
For a server in the server mode, this operand is ignored, if specified.
(v) servexec_retry
User-specified values
Value type |
Range of numerical values |
Can be omitted |
Assumed value when this value is omitted |
---|---|---|---|
Unsigned integer |
0 to 9999 (Units: number of times) |
Yes |
2 |
Description
Specifies the number of times server startup processing is to be re-executed if a server failure is detected on the active server in the monitor mode.
Specify the same value for all hosts in the hot-standby configuration.
If HA Monitor detects a server failure on the active server, it retries execution of the server start command for the server specified in the name operand as many times as specified in this operand. If the server completes startup processing successfully within the retry count, HA Monitor treats it as a normal operation and continues processing. When the number of server failures detected exceeds the specified count, HA Monitor performs automatic planned hot standby processing.
If the server start command returns a return value other than 0 and server restart fails, hot-standby switchover does not occur. If you want to perform hot standby processing when a restart fails, create the start command in such a manner that it always returns a value of 0 and use the monitoring command to detect server start errors.
If 0 is specified in this operand, HA Monitor performs automatic planned hot standby processing immediately after detecting completion of the server monitoring command without the server restarting.
If 9999 is specified in this operand, HA Monitor retries until the server startup processing is completed or the active server is terminated. In such a case, the server might not start successfully for reasons such as that the path specified in the name operand in the server environment definition cannot be found or the required execution permissions are not granted. Therefore, if you specify 9999, it is especially important that you ensure that the specified environment settings and server start command are correct.
Note that specification of this operand is ignored in the following cases:
-
The server has a program for which the restartcommand operand is specified in the monitor-mode program environment definition.
-
server is specified in the acttype operand.
-
No server monitoring command is used (no server monitoring command is specified for the ptrlcmd_ex and patrolcommand operands).
If the switch_retry operand is specified and the target server has already started, the switch_retry operand takes effect. If the server becomes ready for hot standby processing, the servexec_retry operand takes effect.
(w) retry_stable
User-specified values
Value type |
Range of numerical values |
Can be omitted |
Assumed value when this value is omitted |
---|---|---|---|
Unsigned integer |
60 to 3600 (Units: seconds) |
Yes |
60 |
Description
Specifies a monitoring interval (in seconds) for determining whether operation of a server in the monitor mode has become stable after the server has been restarted successfully.
Specify the same value for all hosts in the hot-standby configuration.
Note that specification of this operand is ignored if the ptrlcmd_ex operand is specified.
If no failure is detected within the time specified in this operand after the server restarted, HA Monitor determines that the server's operation has become stable and resets the restart count to 0. After a successful restart, HA Monitor starts monitoring for stable operation beginning at second 0.
If a failure occurs within the time specified in this operand after the server was restarted, the restart count is updated and then the server is restarted again. If the updated restart count exceeds the servexec_retry operand value specified in the server environment definition, HA Monitor performs hot standby processing without restarting the server.
(x) sby_actcommand
User-specified values
Value type |
Range of numerical values |
Can be omitted |
Assumed value when this value is omitted |
---|---|---|---|
Path name having 1 to 1000 characters |
-- (Units: --) |
Yes |
-- |
Description
This operand specifies the absolute path of the command that is to be executed when the standby server starts in the case where HA Monitor performs monitoring in the standby system.
For a server in the server mode, this operand is ignored, if specified.
You can include arguments in the value specified for this operand by enclosing the value in double quotation marks ("). A double quotation mark is not counted as a character. Specify the same value for all hosts in the hot-standby configuration.
(y) sby_termcommand
User-specified values
Value type |
Range of numerical values |
Can be omitted |
Assumed value when this value is omitted |
---|---|---|---|
Path name having 1 to 1000 characters |
-- (Units: --) |
Yes |
-- |
Description
This operand specifies the absolute path of the command that is to be executed when the standby server terminates in the case where HA Monitor performs monitoring in the standby system.
For a server in the server mode, this operand is ignored, if specified.
You can include spaces in the absolute path name by enclosing the entire name in double quotation marks ("). By using spaces, you can specify arguments for the command. A double quotation mark is not counted as a character. If you use a specification that includes spaces, the first character string before the first space is recognized as the path name of the server termination command. The character strings after the first space are recognized as arguments. An argument passed by HA Monitor (-e or -c) is added to the server termination command as the last argument when the command is executed. For details about the arguments that are passed by HA Monitor, see 6.14.2 Creating a command to be executed when the standby server stops.
(z) sby_ptrlcmd_ex
User-specified values
Value type |
Range of numerical values |
Can be omitted |
Assumed value when this value is omitted |
---|---|---|---|
Path name having 1 to 1000 characters |
-- (Units: --) |
Yes |
--# |
- #
-
If this operand is omitted, HA Monitor does not monitor the standby server.
Description
This operand specifies the absolute path of the command that is to be executed for a monitoring purpose in the case where HA Monitor performs monitoring in the standby system.
You can monitor, for example, processes started by sby_actcommand and resources.
You must specify the same value for all hosts in the hot-standby switchover configuration.
For a server in the server mode, this operand is ignored, if specified.
Note that you can include arguments in the path name by enclosing it in double quotation marks (").
If you specify this operand, you must also specify the sby_ptrlcmd_ex_inter operand. If necessary, specify the sby_ptrlcmd_ex_tmout operand.
(aa) sby_ptrlcmd_ex_inter
User-specified values
Value type |
Range of numerical values |
Can be omitted |
Assumed value when this value is omitted |
---|---|---|---|
Unsigned integer |
3 to 180 (Units: seconds) |
Yes# |
-- |
- #
-
You must specify this operand if you specify the sby_ptrlcmd_ex operand.
Description
This operand specifies the interval at which the monitoring command is to be executed if sby_ptrlcmd_ex is specified. Make sure that the interval you specify is longer than the time period from the start to end of the monitoring command.
You must specify the same value for all hosts in the hot-standby switchover configuration.
For a server in the server mode, this operand is ignored, if specified.
(ab) sby_ptrlcmd_ex_retry
User-specified values
Value type |
Range of numerical values |
Can be omitted |
Assumed value when this value is omitted |
---|---|---|---|
Unsigned integer |
0 to 19 (Units: number of times) |
Yes |
2 |
Description
This operand specifies the number of times HA Monitor can retry execution of the monitoring command specified for sby_ptrlcmd_ex if the command fails (with the EXIT code 1 to 9).
You must specify the same value for all hosts in the hot-standby switchover configuration.
If an attempt to execute the command fails the specified number of times in succession, HA Monitor assumes that a failure occurred and operates as specified by the sby_servexec_retry operand.
(ac) sby_ptrlcmd_ex_tmout
User-specified values
Value type |
Range of numerical values |
Can be omitted |
Assumed value when this value is omitted |
---|---|---|---|
Unsigned integer |
5 to 3600 (Units: seconds) |
Yes |
-- |
Description
This operand specifies the timeout value for execution of the monitoring command specified for sby_ptrlcmd_ex. Make sure that the timeout value you specify is equal to or longer than the time period from the start to end of the monitoring command. If you specify a value shorter than the time period from the start to end of the monitoring command, monitoring cannot be performed appropriately because HA Monitor might perform a false detection.
If the command times out, HA Monitor assumes that a server failure occurred, and operates as specified by the sby_servexec_retry operand.
You must specify the same value for all hosts in the hot-standby switchover configuration.
(ad) sby_start_timeout
User-specified values
Value type |
Range of numerical values |
Can be omitted |
Assumed value when this value is omitted |
---|---|---|---|
Unsigned integer |
60 to 3600 (Units: seconds) |
Yes |
-- |
Description
This operand specifies the timeout value for the command specified for sby_actcommand.
You must specify the same value for all hosts in the hot-standby switchover configuration.
Note that HA Monitor does not monitor the timeout for the command started when the standby server's startup processing is retried according to the specification of the sby_servexec_retry operand.
If a timeout occurs, HA Monitor forcibly terminates the command specified for sby_actcommand and cancels the standby server's startup processing. After the monitoring-target command times out, the command specified for sby_termcommand starts. Make sure that the timeout value you specify is equal to the time period from the start to end of the command specified for sby_actcommand.
For a server in the server mode, this operand is ignored, if specified.
(ae) sby_stop_timeout
User-specified values
Value type |
Range of numerical values |
Can be omitted |
Assumed value when this value is omitted |
---|---|---|---|
Unsigned integer |
60 to 3600 (Units: seconds) |
Yes |
-- |
Description
This operand specifies the timeout value for the command specified for sby_termcommand.
You must specify the same value for all hosts in the hot-standby switchover configuration.
HA Monitor does not monitor the timeout for the following command:
-
The command that is specified for sby_termcommand to be executed when the standby server is restarted according to the specification of the sby_servexec_retry operand
Make sure that the timeout value you specify is equal to the time period from the start to end of the command specified for sby_termcommand.
If a timeout occurs, HA Monitor forcibly terminates the command specified for sby_termcommand.
For a server in the server mode, this operand is ignored, if specified.
(af) sby_servexec_retry
User-specified values
Value type |
Range of numerical values |
Can be omitted |
Assumed value when this value is omitted |
---|---|---|---|
Unsigned integer |
0 to 9999 (Units: number of times) |
Yes |
0 |
Description
This operand specifies the number of times the standby server's startup processing is retried if a failure is detected by the monitoring command specified for sby_ptrlcmd_ex.
You must specify the same value for all hosts in the hot-standby switchover configuration.
Upon detecting a failure, HA Monitor retries execution of the command specified for the sby_actcommand operand as many times as specified in this operand. If startup processing of the standby server is completed successfully within the retry count, HA Monitor treats it as a normal operation and continues processing. If a failure is still detected after command execution is retried the specified number of times, HA Monitor terminates the standby server.
In the following cases, HA Monitor does not restart the standby server and terminates the standby server immediately upon detecting termination of the monitoring command:
-
This operand is omitted.
-
0 is specified for this operand.
-
No command is specified for the sby_actcommand operand.
If 9999 is specified in this operand, HA Monitor retries until the standby server startup processing is completed or the standby server is terminated. In such a case, the standby server might not start successfully for reasons such as that the path specified in the sby_actcommand operand in the server environment definition cannot be found or the required execution permissions are not granted. Therefore, it is especially important that you ensure that the specified environment settings and the command specified for the sby_actcommand operand are correct.
For a server in the server mode, this operand is ignored, if specified.
(ag) switchtype
User-specified values
Value type |
Range of numerical values |
Can be omitted |
Assumed value when this value is omitted |
---|---|---|---|
switch|restart|manual |
-- (Units: --) |
Yes |
switch |
Description
Specifies the processing to be performed when a server failure is detected on the server in the server mode.
Specify the same value for all hosts in the hot-standby configuration. For a server in the monitor mode, this operand is ignored, if specified.
-
switch: Terminates the active server and performs hot standby processing.
-
restart: Places the active server in the active-server-restart wait state and waits until the active server is restarted automatically. HA Monitor does not perform hot standby processing at this point. If restart of the active server fails, HA Monitor retries the restart processing as many times as specified for the server (retry count). If the restart is not successful after the specified number of retries has been attempted, HA Monitor starts hot standby processing.
When you specify restart, you can also specify an active-server-restart monitoring time by separating the values with a colon (:). If the active server has not been restarted successfully after the specified restart monitoring time has been reached, HA Monitor starts hot standby processing. For the restart monitoring time, you can specify a value in the range from 60 to 3,600 seconds. If you omit the restart monitoring time, HA Monitor does not monitor for restart of the active server.
-
manual: Places the active server in the active-server-restart wait state and waits until the active server is restarted automatically. HA Monitor does not perform hot standby processing at this point. If restart of the active server fails, HA Monitor retries the restart processing as many times as specified for the server (retry count). If the restart is not successful after the specified number of retries has been attempted, HA Monitor terminates the active server, terminates the corresponding standby server, restarts the standby server, then places the standby server in the active server start wait state.
If HA Monitor detects a slowdown on the server when manual is specified, HA Monitor continues monitoring the server without performing any processing.
(ah) switch_error
User-specified values
Value type |
Range of numerical values |
Can be omitted |
Assumed value when this value is omitted |
---|---|---|---|
retry|ignore |
-- (Units: --) |
Yes |
ignore |
Description
Specifies whether startup of a server is to be retried automatically as the active system if a shared resource connection error occurs while the target active server is starting or the active server fails to start and there is no available switchover host.
You must specify the same value for all hosts in the hot-standby configuration.
-
retry: Automatically retries startup of server as the active system.
-
ignore: Cancels server startup.
To have HA Monitor retry startup processing in the server mode in the event of a hot standby error, you must specify the following settings on the server:
-
HiRDB
Specify retry in the pd_ha_switch_error operand in the HiRDB system definition (pdsys or pdutsys).
-
TP1/Server Base
Specify a desired retry count in the ha_switch_error_retry_count operand in the TP1/Server Base system environment definition.
For a server in the monitor mode or a resource server, you can specify the start retry count in the switch_retry operand.
The following table describes the definition of the switch_error and switch_retry operands and the HA Monitor processing.
Server type |
Definition of switch_error |
Definition of switch_retry |
HA Monitor processing |
---|---|---|---|
Server mode |
Omitted or ignore |
-- |
|
retry |
-- |
|
|
Monitor mode or resource server |
Omitted or ignore |
Omitted |
|
Specified |
Issues the KAMN514-E message and stops moncheck. |
||
retry |
Omitted or specified |
|
- Legend:
-
--: Undefined (ignored if defined)
The switch_error operand is ignored, if specified, for a server that contains a program for which the restartcommand operand is specified in the monitor-mode program environment definition.
For details about start retries during hot standby processing, see 3.2.7 Performing start retries during hot standby processing.
(ai) switch_retry
User-specified values
Value type |
Range of numerical values |
Can be omitted |
Assumed value when this value is omitted |
---|---|---|---|
Unsigned integer |
1 to 9999 (Units: number of times) |
Yes |
2 |
Description
Specifies the number of times startup of a server in the monitor mode or a resource server is to be attempted automatically when a shared resource connection error occurs while the target active server is starting or the active server fails to start and there is no available switchover host.
You must specify the same value for all hosts in the hot-standby configuration.
If 9999 is specified in this operand, HA Monitor attempts to restart the server indefinitely until the server startup processing is completed or the start retry state is released forcibly.
To perform start retries in the event of a hot standby error, specify retry in the switch_error operand.
(aj) switchbyfail
User-specified values
Value type |
Range of numerical values |
Can be omitted |
Assumed value when this value is omitted |
---|---|---|---|
1 to 32 alphanumeric characters |
-- (Units: --) |
Yes |
-- |
Description
Specifies the names of the LAN interfaces that are to be monitored.
If a failure occurs on any LAN interface specified in this operand, HA Monitor performs hot standby processing on the active server. If an alias IP address is assigned to a LAN interface that is to be monitored, specify the alias IP address's interface name without the index number.
You can specify a maximum of 32 LAN interface names by delimiting them with the colon. When you specify this operand, you must create a LAN monitor definition file for each LAN interface. If a server for which this operand is specified is grouped with other servers and a failure occurs on a specified LAN interface, all servers in the group will be switched over.
For details about the LAN monitor definition file, see 6.20.1 Settings in the files required for monitoring a LAN.
(ak) disk_ptrl
User-specified values
Value type |
Range of numerical values |
Can be omitted |
Assumed value when this value is omitted |
---|---|---|---|
use|nouse |
-- (Units: --) |
Yes |
nouse |
Description
Specifies whether to monitor the disks for business use on the active server.
Specify the same values for all hosts in the hot-standby configuration.
-
use: HA Monitor monitors the disks for business use.
-
nouse: HA Monitor does not monitor the disks for business use.
If you specify use for this operand, you must set up the monitoring definition file for the disk for business use. For details about how to specify the settings in a monitoring definition file for the disk for business use, see (2) Settings in the files required for monitoring disks for business use in 6.20.2 Settings in the files required for disk monitoring.
If necessary, also specify the following operands.
- Server environment definition:
-
-
disk_ptrl_inter
-
disk_ptrl_retry
-
disk_ptrl_act
-
- HA Monitor environment settings:
-
-
disk_log_size
-
(al) disk_ptrl_inter
User-specified values
Value type |
Range of numerical values |
Can be omitted |
Assumed value when this value is omitted |
---|---|---|---|
Unsigned integer |
3 to 3600 (Units: seconds) |
Yes |
120 |
Description
Specifies the interval at which HA Monitor checks a disk for business use on the active server.
Specify the same values for all hosts in the hot-standby configuration.
Determine the value to be specified for this operand taking into account the following items specified for the disk_ptrl_retry operand in the server environment definition:
-
Check retry interval for a disk for business use
-
Check retry count for a disk for business use
Specify a value that meets the following condition:
check-interval-for-disk-for-business-use ≥ check-retry-interval-for-disk-for-business-use × (check-retry-count-for-disk-for-business-use + 1)
(am) disk_ptrl_retry
User-specified values
Value type |
Range of numerical values |
Can be omitted |
Assumed value when this value is omitted |
---|---|---|---|
check-interval-for-disk-for-business-use: Unsigned integer |
1 to 3600 (Units: seconds) |
Yes |
60 |
check-retry-count-for-disk-for-business-use: Unsigned integer |
0 to 60 (Units: number of times) |
Yes |
1 |
Description
Use this operand to specify the interval at which HA Monitor retries checking a disk for business use on the active server and the number of times HA Monitor can retry checking it if access to the disk fails. Specify the retry interval and count with an intervening colon (:). If the number of times access to a disk for business use fails exceeds the specified count, HA Monitor judges that the disk has failed.
Specify the same values for all hosts in the hot-standby configuration.
-
Check retry interval for the disks for business use
Specify the interval at which HA Monitor retries checking a disk for business use. Make sure that the value you specify is equal to or larger than the disk I/O timeout value. The disk I/O timeout value means the time for which an application that requests an I/O operation can wait before receiving a response. This value is determined from the driver settings, OS kernel parameters, and multipath software settings.
-
Check retry count for a disk for business use
Specify the number of times HA Monitor can retry checking a disk for business use. If you specify 0, HA Monitor does not retry checking a disk for business use.
(an) disk_ptrl_act
User-specified values
Value type |
Range of numerical values |
Can be omitted |
Assumed value when this value is omitted |
---|---|---|---|
switch|pairdown|noswitch |
-- (Units: --) |
Yes |
switch |
Description
Specifies the action that HA Monitor takes when a failure occurs on a disk for business use on the active server.
Specify the same values for all hosts in the hot-standby configuration.
-
switch: HA Monitor performs planned hot-standby switchover for servers.
-
pairdown: HA Monitor causes host pair shutdown to perform hot-standby switchover as a host failure.
-
noswitch: HA Monitor does not perform hot-standby switchover and continues monitoring of the disk for business use.
(ao) switch_judge
User-specified values
Value type |
Range of numerical values |
Can be omitted |
Assumed value when this value is omitted |
---|---|---|---|
1 to 32 alphanumeric characters |
-- (Units: --) |
Yes |
-- |
Description
Specifies the name of a LAN interface whose availability is to be checked during hot standby processing resulting from a host failure.
This operand can be specified only when the fence_lan operand is specified in the HA Monitor environment settings.
HA Monitor controls in such a manner that the server that can establish communication via the LAN interface specified in this operand becomes the active server. If an alias IP address has been assigned to the target LAN interface, specify the interface name without the index number for the alias IP address.
If you group servers, specify this operand only for the server at the top of the hierarchy or a resource server. If there are multiple servers at the top in the group, specify the same LAN interface name for all such servers.
When you specify this operand, you must create a LAN monitor definition file for the LAN interface.
For details about the function for controlling hot standby based on the availability of LAN communications, see 2.3.7 Function for controlling hot standby based on the availability of LAN communications.
For details about LAN monitor definition files, see 6.20.1 Settings in the files required for monitoring a LAN.
(ap) pairdown
User-specified values
Value type |
Range of numerical values |
Can be omitted |
Assumed value when this value is omitted |
---|---|---|---|
use|nouse |
-- (Units: --) |
Yes |
nouse |
Description
Specifies whether the pair shutdown function that terminates HA Monitor when a server failure is detected is to be used.
Specify the same value for all hosts in the hot-standby configuration. For a server in the monitor mode, specification of this operand is ignored unless the ptrlcmd_ex operand is specified.
-
use: Uses the pair shutdown function. If a failure occurs on the active server, pair shutdown is performed. If TP1/Server Base or HiRDB is used as the server, you can add the following value after use. To do so, use a colon (:) as a separator.
serv_slow
Performs pair shutdown only when server slowdown of the active server is detected. Pair shutdown is not performed when the server itself detects a failure. You can specify serv_slow only if switch is specified for the switchtype operand. In the case where B is specified for the server_type operand or where the ptrlcmd_ex operand is specified, if you specify serv_slow, HA monitor outputs KAMN163-E and stops parsing the definitions.
-
nouse: Does not use the pair shutdown function.
(aq) uoc_neck
User-specified values
Value type |
Range of numerical values |
Can be omitted |
Assumed value when this value is omitted |
---|---|---|---|
use|nouse |
-- (Units: --) |
Yes |
nouse |
Description
Specifies whether the termination code is to be checked for the user command that is executed when server startup processing or hot standby processing is started.
Specify the same value for all hosts in the hot-standby configuration.
-
use: Cancels server startup processing if the user command's termination code is not 0.
-
nouse: Continues server startup processing without checking the user command's termination code.
Note that termination code checking is not performed when no user command is executed at the time that server startup processing or hot standby processing is started.
(ar) uoc_pairdown
User-specified values
Value type |
Range of numerical values |
Can be omitted |
Assumed value when this value is omitted |
---|---|---|---|
use|nouse |
-- (Units: --) |
Yes |
nouse |
Description
Specifies whether the pair shutdown function is to be used based on the termination code of the user command that is executed at the following time:
-
When hot standby processing begins due to a server failure
(server type: online; server status parameter: -a; start or end parameter: start)
-
When planned hot standby processing begins
(server type: online; server status parameter: -w; start or end parameter: start)
You must specify the same value for all hosts in the hot-standby configuration.
-
use: Uses the pair shutdown function that also terminates HA Monitor when the user command's termination code is not 0.
-
nouse: Resumes server hot standby processing without checking the user command's termination code.
Note that if use is specified, pair shutdown is also performed in the following cases:
-
The user command cannot be executed.
-
Processing is canceled according to the specification of the dev_timelimit operand in the server environment definition.
-
Processing is canceled according to the specification of the dev_offlimit operand in the server environment definition.
For details about host pair shutdown performed for each host in the event of a user command error, see (2) Host pair shutdown performed for each host in the event of a user command error.
(as) term_pairdown
User-specified values
Value type |
Range of numerical values |
Can be omitted |
Assumed value when this value is omitted |
---|---|---|---|
use|nouse |
-- (Units: --) |
Yes |
nouse |
Description
This operand specifies whether to use the pair shutdown function if the server termination command returns a termination code other than 0.
You must specify the same value for all hosts in the hot-standby switchover configuration.
-
use: Uses the pair shutdown function, which also terminates HA Monitor, if the server termination command returns a termination code other than 0.
-
nouse: Resumes server hot-standby switchover without checking the termination code of the server termination command.
For details about host pair shutdown performed for each host if the server termination command fails, see (3) Host pair shutdown for each host in the event of a server termination command error.
For a server in the server mode, this operand is ignored, if specified.
(at) lan_updown
User-specified values
Value type |
Range of numerical values |
Can be omitted |
Assumed value when this value is omitted |
---|---|---|---|
use|nouse |
-- (Units: --) |
Yes |
--# |
#: If this operand is omitted but server-alias-name.up and server-alias-name.down files exist under the directory for HA Monitor environment settings, HA Monitor uses those files as LAN status settings files anyway.
Description
Specifies whether LAN status settings files are to be used.
Specify the same value for all hosts in the hot-standby configuration.
-
use: Uses LAN status settings files (server-alias-name.up and server-alias-name.down files).
-
nouse: Does not use LAN status settings files.
(au) ip_neck
User-specified values
Value type |
Range of numerical values |
Can be omitted |
Assumed value when this value is omitted |
---|---|---|---|
use|nouse |
-- (Units: --) |
Yes |
nouse |
Description
Specifies whether the server-alias-name.up file's termination code is to be checked when the file is executed during server startup processing or hot standby processing.
Specify the same value for all hosts in the hot-standby configuration.
-
use: Cancels server startup processing if the server-alias-name.up file's termination code is not 0. If you specify this value, specify LAN status settings files by referencing the example of a server-alias-name.up file provided in 6.12.2 Specifying LAN status settings files.
-
nouse: Continues server startup processing without checking the termination code from the server-alias-name.up file. When you specify this value, specify LAN status settings files by referencing the example of a server-alias-name.up file provided in 6.12.2 Specifying LAN status settings files.
(av) disk
User-specified values
Value type |
Range of numerical values |
Can be omitted |
Assumed value when this value is omitted |
---|---|---|---|
Path name having 1 to 256 characters |
-- (Units: --) |
Yes |
--# |
#: If this operand is omitted, HA Monitor does not perform switchover of volume groups.
Description
Specifies the absolute path names of volume groups defined on a shared disk that are to be switched over when a shared disk is used.
You can specify a maximum of 3,000 path names. Specify the same values for all hosts in the hot-standby configuration.
Notes
A volume group created on a disk device included in a control group of HA Booster cannot be specified for the disk operand. If you inadvertently specify such a volume group, server startup or hot-standby switchover might fail. For a control group of HA Booster, specify the control group ID for the hab_gid operand.
(aw) vg_on_opt
User-specified values
Value type |
Range of numerical values |
Can be omitted |
Assumed value when this value is omitted |
---|---|---|---|
Character string having 1 to 256 characters |
-- (Units: --) |
Yes |
-aΔy |
Description
Specifies for the vgchange command run-time options that are to be used when volume groups to be switched over are connected.
Specify the same values for all hosts in the hot-standby configuration.
If this operand is omitted, the -aΔy option is assumed.
If an option contains a space (Δ), enclose the entire option in double quotation marks (").
The options in the vg_on_opt operand must be specified so that they correspond in order and number to the volume groups specified in the disk operand. You must specify as many sets of options as there are values in the disk operand. To omit an option for a volume group at any intermediate location, specify a colon (:) without specifying any value.
The following shows an example of specifying vgchange command run-time options:
disk /dev/vg01 :/dev/vg02 , ↓ vg_on_opt "-aΔyΔ-qΔn" : ,
In the preceding example, for the /dev/vg01 volume group specified by the disk operand, the vgchange command is executed with the -aΔyΔ-qΔn option. For the /dev/vg02 volume group, the vgchange command is executed with the -aΔy option (default option).
If the vg_on_opt operand is omitted, the default value is assumed as the option for all volume groups.
(ax) vg_neck
User-specified values
Value type |
Range of numerical values |
Can be omitted |
Assumed value when this value is omitted |
---|---|---|---|
use|nouse |
-- (Units: --) |
Yes |
nouse |
Description
Specifies for each volume group whether the server startup processing is to be cancelled if the volume group's connection processing fails during server startup processing or hot standby processing.
Specify the same value for all hosts in the hot-standby configuration.
-
use: Cancels server startup processing if connection fails.
-
nouse: Continues server startup processing even if connection fails.
As with the vg_on_opt operand, specify the options in the vg_neck operand so that they corresponding in order and number to the volume groups specified in the disk operand. You must specify as many options as there are values in the disk operand. To omit the option for a volume group at any intermediate location, specify a colon (:) without specifying any value (such as vg_neck use::nouse,). If the vg_neck operand is omitted, nouse is assumed for all volume groups.
(ay) fs_name
User-specified values
Value type |
Range of numerical values |
Can be omitted |
Assumed value when this value is omitted |
---|---|---|---|
Path name having 1 to 256 characters |
-- (Units: --) |
Yes |
--# |
#: If this operand is omitted, HA Monitor does not switch over file systems on the shared disk.
Description
Use this operand to specify the path names of file systems.
If LVM is used, specify the absolute path names of logical volumes in the "/dev/volume-group-name/logical-volume-name" format.
If LVM is not used, specify the universally unique identifiers (UUIDs) of file systems in the "/dev/disk/by-uuid/uuid" format.
You can use the lsblk -f command to check the UUID of a file system from the name of the device on which the file system was created. The following shows an example of checking the UUID of a file system that has been created on /dev/sda.
# lsblk -f NAME FSTYPE LABEL UUID sda xfs c73bf13f-89a9-49bc-9d8d-7c94e5458389
You can specify a maximum of 3,000 file system path names. Specify the same values for all hosts in the hot-standby configuration.
When you specify the fs_name operand, you must also specify the fs_mount_dir operand.
(az) fs_mount_dir
User-specified values
Value type |
Range of numerical values |
Can be omitted |
Assumed value when this value is omitted |
---|---|---|---|
Path name having 1 to 256 characters |
-- (Units: --) |
Yes |
-- |
Description
Specifies the absolute path names of the directories where the file systems to be switched over are to be mounted when file systems on the shared disk are used.
Specify the same values for all hosts in the hot-standby configuration.
You must specify this operand if you have specified the fs_name operand.
The directories in the fs_mount_dir operand must be specified so they correspond in order and number to the logical volumes specified in the fs_name operand. Make sure that you specify the same number of mount target directories as there are path names in the fs_name operand. The following shows an example of the fs_name and fs_mount_dir operands:
fs_name /dev/vg01/lvol1:/dev/vg02/lvol1, ↓ fs_mount_dir /home1 :/home2,
In this example, /dev/vg01/lvol1 in the fs_name operand corresponds to /home1 in the fs_mount_dir operand, and /dev/vg02/lvol1 in the fs_name operand corresponds to /home2 in the fs_mount_dir operand.
File systems specified in this operand must not be set to be mounted automatically during system startup.
(ba) fs_mount_opt
User-specified values
Value type |
Range of numerical values |
Can be omitted |
Assumed value when this value is omitted |
---|---|---|---|
Character string having 1 to 256 characters |
-- (Units: --) |
Yes |
--# |
#: If this operand is omitted, the run-time options are omitted from the mount command for the file system.
Description
Specifies mount command run-time options for the file systems that are to be switched over (when file systems on a shared disk are used).
Specify the same values for all hosts in the hot-standby configuration. If an option contains a space, comma, colon, semicolon, forward slash, or asterisk, enclose the option in double quotation marks (").
The options in the fs_mount_opt operand must be specified so that they correspond in order and number to the file systems specified in the fs_name and fs_mount_dir operands. You must specify as many sets of options as there are values in the fs_name and fs_mount_dir operands. To omit an option for a file system at any intermediate location, specify a colon (:) without specifying any value. The following shows an example of fs_name, fs_mount_dir, and fs_mount_opt operands:
fs_name /dev/vg01/lvol1:/dev/vg02/lvol1, ↓ fs_mount_dir /home1 :/home2, ↓ fs_mount_opt "-o rw" :,
In this example, -o rw in the fs_mount_opt operand corresponds to /dev/vg01/lvol1 in the fs_name operand and to /home1 in the fs_mount_dir operand. No fs_mount_opt operand value is set for /dev/vg02/lvol1 in the fs_name operand and for /home2 in the fs_mount_dir operand.
(bb) fs_umount_retry
User-specified values
Value type |
Range of numerical values |
Can be omitted |
Assumed value when this value is omitted |
---|---|---|---|
Unsigned integer |
0 to 9999 (Units: number of times) |
Yes |
10 |
Description
Specifies a retry count to be applied if an unmount operation fails when file systems are disconnected.
Specify the same value for all hosts in the hot-standby configuration.
If 0 is specified, an unmount operation is not retried. If 9999 is specified, the unmount operation is retried until it is successful. The retry interval is one second.
(bc) fs_neck
User-specified values
Value type |
Range of numerical values |
Can be omitted |
Assumed value when this value is omitted |
---|---|---|---|
use|nouse |
-- (Units: --) |
Yes |
nouse |
Description
Specifies for each file system whether the server startup processing is to be cancelled if the file system's connection processing fails during server start or hot standby processing.
Specify the same value for all hosts in the hot-standby configuration.
-
use: Cancels server startup processing if connection fails.
-
nouse: Continues server startup processing even if connection fails.
As with the fs_mount_opt operand, specify the options in the fs_neck operand so that they correspond in order and number to the file systems specified in the fs_name operand. You must specify as many options as there are values in the fs_name operand. You can omit options in a list of colon-separated values. To omit options, only omit specifying the values and do not omit colons (:) in a format such as fs_neck use::nouse,. If the fs_neck operand is omitted, HA Monitor assumes that nouse is specified for all file systems.
In RHEL7 or a later version, if XFS is used as the file system of the shared disk, HA Monitor uses the xfs_repair command to check the consistency of the file system. Before the consistency of the file system is checked, mount and unmount processing is performed. If unmount processing fails, HA Monitor stops starting the server regardless of the values specified for this operand.
(bd) scsi_device
User-specified values
Value type |
Range of numerical values |
Can be omitted |
Assumed value when this value is omitted |
---|---|---|---|
Path name having 1 to 384 characters |
-- (Units: --) |
Yes |
-- |
Description
When SCSI reservation for shared disk is used, this operand specifies the absolute path of a device subject to SCSI reservation. For notes on specifying the SCSI reservation settings, see 6.4.6 Notes about SCSI reservation for shared disk.
Express a path name as 1 to 384 alphabetic characters, numeric characters, forward slashes (/), hyphens (-), underscores (_), and periods (.). You can specify a maximum of 3,000 path names.
Specify a path name value as follows:
-
If HFC-PCM or EMPD is used with a single-path configuration in a VMware ESXi-based virtualization environment
Specify the absolute path of the symbolic link under /dev/disk/by-id/ that corresponds to the physical device used for the shared disk. Specify the same names in the same order on all hosts in the hot-standby configuration. For details about the symbolic link, see (a) If HFC-PCM is used with a single-path configuration in a VMware ESXi-based virtualization environment.
Note that this operand is not needed for a redundant configuration using DMMP even when a VMware ESXi-based virtualization environment is used. Specify the dmmp_device operand.
-
Redundant configuration using HDLM
Specify the absolute path of the HDLM logical device that is used as the shared disk. Specify the same absolute paths in the same order on all hosts in the hot-standby configuration. If the same disk is used for all hosts in the hot-standby configuration, you can specify different names. For details about the HDLM logical devices that make up a shared disk, see (b) Redundant configuration using multipath software (HDLM).
When you specify this operand, you must specify use in the fence_scsi operand in the HA Monitor environment settings. The server's startup processing will fail if the specified value is neither a symbolic link nor HDLM logical device.
In this case, before you restart the server, correct the operand value and then execute the definition check command (moncheck command) to make sure that neither the KAMN113-E message nor the KAMN723-E message is issued.
(be) dmmp_device
User-specified values
Value type |
Range of numerical values |
Can be omitted |
Assumed value when this value is omitted |
---|---|---|---|
Path name having 1 to 256 characters |
-- (Units: --) |
Yes |
-- |
Description
When SCSI reservation for shared disk is used in a multi-path configuration, specifies a device subject to SCSI reservation, expressed as the absolute path of the logical device of DMMP. For notes on specifying the SCSI reservation settings, see 6.4.6 Notes about SCSI reservation for shared disk.
Specify the same names in the same order on all hosts in the hot-standby configuration. For details about the logical device name of DMMP, see (2) dmmp_device operand value (redundant configuration using multipath software (DMMP)).
Express a path name as 1 to 256 alphabetic characters, numeric characters, forward slashes (/), hyphens (-), underscores (_), and periods (.). You can specify a maximum of 3,000 path names.
When you specify this operand, you must specify use in the fence_scsi operand in the HA Monitor environment settings.
The server's startup processing will fail in the following cases:
-
use is not specified in the fence_scsi operand in the HA Monitor environment settings.
-
The specified name is not the logical device name of DMMP.
In this case, before you restart the server, correct the operand value and then execute the definition check command (moncheck command) to make sure that neither the KAMN113-E message nor the KAMN723-E message is issued.
(bf) deviceoff_order
User-specified values
Value type |
Range of numerical values |
Can be omitted |
Assumed value when this value is omitted |
---|---|---|---|
order|reverse |
-- (Units: --) |
Yes |
order |
Description
Specifies the order in which shared resources are to be disconnected.
Specify the same value for all hosts in the hot-standby configuration. In contrast to the deviceoff_order operand in the HA Monitor environment settings, this operand is applicable only to the specified server.
If the deviceoff_order operand is specified in the HA Monitor environment settings, this operand is ignored, in which case the specified HA Monitor environment setting is effective. If you want to change the order in which shared resources are to be disconnected on a per-server basis, specify this operand and do not specify the deviceoff_order operand in the HA Monitor environment settings.
For details about the order in which shared resources are disconnected, see 4.8.7 Processing flow for disconnecting shared resources in the reverse order from which they were connected.
-
order: Disconnects the shared resources in the same order as when they were connected.
-
reverse: Disconnects the shared resources in the reverse order from which they were connected.
(bg) dev_timelimit
User-specified values
Value type |
Range of numerical values |
Can be omitted |
Assumed value when this value is omitted |
---|---|---|---|
Unsigned integer |
60 to 3600 (Units: seconds) |
Yes |
-- |
Description
Specifies a timeout value for shared resource connection and disconnection processing for the server during hot standby processing.
We recommend that you use the dev_offlimit and dev_onlimit operands to specify the timeout values for disconnection and connection of resources shared by servers at the time of hot-standby switchover. We do not recommend that you use the dev_timelimit operand for that purpose. This is because the system might behave as follows if the dev_timelimit operand is used:
-
If too small a timeout value is specified, hot-standby switchover might fail.
-
If too large a timeout value is specified and disconnection of shared resources takes time, hot-standby switchover also takes time.
Specify the same value for all hosts in the hot-standby configuration.
If too small a timeout value is specified and restoration or mounting of a file system takes time at the hot-standby switchover destination, switchover might fail. To avoid this, specify a sufficiently large value such as a few minutes or the maximum value (3600).
If a timeout occurs, HA Monitor treats the shared resource connection or disconnection processing as having failed and cancels the processing. It then continues the server start or termination processing. If this operand is omitted, HA Monitor does not monitor the timer and places server start or termination processing on hold until connection or disconnection processing is completed.
You cannot use the dev_timelimit operand to specify different timeout values for shared resource connection processing and for shared resource disconnection processing. Specify a timeout value for the following operands:
-
The dev_onlimit operand in the server environment definition (timeout value for shared resource connection)
-
The dev_offlimit operand in the server environment definition (timeout value for shared resource disconnection)
(bh) dev_offlimit
User-specified values
Value type |
Range of numerical values |
Can be omitted |
Assumed value when this value is omitted |
---|---|---|---|
Unsigned integer |
1 to 3600 (Units: seconds) |
Yes |
-- |
Description
Specifies a timeout value for the shared resource disconnection processing performed for the server when hot-standby switchover occurs.
We recommend that you specify this operand (or the dev_timelimit operand) to prevent indefinite waiting for disconnection of shared resources.
You must specify the same value for all hosts in the hot-standby switchover configuration.
If a timeout occurs, HA Monitor treats the processing performed for shared resources as having failed and cancels the processing. HA Monitor then performs SCSI reservation for shared disk and continues hot-standby switchover. If this operand is omitted, HA Monitor does not monitor the timer and places server termination processing on hold until disconnection processing is completed.
You cannot use this operand if you have specified the dev_timelimit operand.
(bi) dev_onlimit
User-specified values
Value type |
Range of numerical values |
Can be omitted |
Assumed value when this value is omitted |
---|---|---|---|
Unsigned integer |
60 to 3600 (Units: seconds) |
Yes |
-- |
Description
Specifies a timeout value for the shared resource connection processing performed for the server when hot-standby switchover occurs.
We recommend that you specify this operand (or the dev_timelimit operand) to prevent indefinite waiting for connection of shared resources.
You must specify the same value for all hosts in the hot-standby switchover configuration.
If too small a timeout value is specified and restoration or mounting of a file system takes time at the hot-standby switchover destination, switchover might fail. To avoid this, specify a sufficiently large value such as a few minutes or the maximum value (3600).
If a timeout occurs, HA Monitor treats the processing performed for shared resources as having failed and cancels the processing. HA Monitor then continues the server start processing. If this operand is omitted, HA Monitor does not monitor the timer and places server start processing on hold until connection processing is completed.
You cannot use this operand if you have specified the dev_timelimit operand.
(3) resource definition statement
The resource definition statement defines a resource server. The operands of the resource definition statement are explained below. For details about resource servers, see 4.6 Managing a resource server.
The resource definition statement's operands are the same as the operands of the server statement in the server environment definition. For details of the operands, see the server definition statement. This subsection explains only those operands that differ from when they are specified in a server definition statement.
(a) alias
User-specified values
Value type |
Range of numerical values |
Can be omitted |
Assumed value when this value is omitted |
---|---|---|---|
1 to 8 alphanumeric characters |
-- (Units: --) |
No |
-- |
Description
Specifies the alias name of the resource server (server's alias) that is used in HA Monitor commands and in the messages that are issued.
This is the same as the alias operand in the server definition statement.
(b) group
User-specified values
Value type |
Range of numerical values |
Can be omitted |
Assumed value when this value is omitted |
---|---|---|---|
1 to 8 alphanumeric characters |
-- (Units: --) |
No |
-- |
Description
Specifies the name of the server group to be switched over in the batch mode during grouped-system switchover.
Specify the same value for all hosts in the hot-standby configuration.
This operand is required for a resource server. The names exchange, no_exchange, cancel, and no_cancel are ignored, if specified. The other information is the same as for the group operand in the server definition statement.
Note that only one resource server can be defined per group.