25.13.3 Terminating HiRDB (in the server mode)

Organization of this subsection
(1) Standby system switchover facility
(2) Standby-less system switchover (1:1) facility
(3) Standby-less system switchover (effects distributed) facility (terminating both the running system and the standby system)
(4) Standby-less system switchover (effects distributed) facility (terminating only the standby system)

(1) Standby system switchover facility

Table 25-32 lists the methods of terminating HiRDB when using the standby system switchover facility.

Table 25-32 Terminating HiRDB when using the standby system switchover facility

ConditionTermination method
Cluster software used is HA monitorTerminating both the running system and the standby systemUse the pdstop command to terminate the running system HiRDB. The standby system HiRDB will terminate together with the running system. This situation also applies during planned termination or forced termination.
To terminate at the unit level, use the pdstop -u command to terminate the running system unit. The standby system unit will terminate together with the running system unit. This situation also applies when the pdstop -z command is executed.
Terminating only the standby systemUse the HA monitor monsbystp command to terminate only the standby system.
Cluster software other than HA monitor is usedTerminating both the running system and the standby system*Use the pdstop command to terminate the running system HiRDB, then use the hatesbystp command in Hitachi HA Toolkit Extension to terminate HiRDB. Executing the pdstop command does not terminate the standby system HiRDB. This situation also applies to planned termination or forced termination.
To terminate at the unit level, use the pdstop -u command to terminate the running system unit, then use the hatesbystp command in Hitachi HA Toolkit Extension to terminate the standby unit. Executing the pdstop -u command does not terminate the standby system unit. This situation also applies when the pdstop -z command is executed.
Terminating only the standby systemUse the hatesbystp command in Hitachi HA Toolkit Extension to terminate the standby system HiRDB.
*
  • When terminating the running system HiRDB (or unit), be sure to terminate the standby system HiRDB also.
  • Be sure to terminate the standby system HiRDB before restarting the running system HiRDB (or unit).

(2) Standby-less system switchover (1:1) facility

Table 25-33 lists the methods of terminating HiRDB when you use the standby-less system switchover (1:1) facility.

Table 25-33 Terminating HiRDB when using the standby-less system switchover (1:1) facility

ObjectiveCommand to executeRemarks (effect on other unit or alternate portion)
Terminating HiRDBpdstopNo specific operation needs to be performed on the alternate BES. Operation does not change even when alternating units.
Terminating the normal BES unitpdstop -uWaiting status1 of the alternate portion is released.
Terminating the alternate BES unitNormal operationpdstop -uWaiting status of the alternate portion in the alternate BES unit is released.
AlternatingThe alternate portion in the alternate BES unit also stops.
Also, when the normal BES unit is in waiting status,2 the waiting status of the alternate BES unit is released.
Terminating the alternate portion when alternating unitspdstop -uSpecify for the unit identifier the unit identifier of the normal BES unit.
Releasing waiting status of the alternate portionhatesbystp3None
Releasing waiting status of the normal BES unithatesbystp4

Note: Termination at the server level is not possible.

1 The system is switched to the alternate BES unit only if the alternate portion is in waiting status. If the alternate portion is in waiting status, the system status is displayed as SBY in the execution results of the pdls -d ha command.

2 The system will be switched back to the normal BES unit only if the normal BES unit is in waiting status (that is, it will not be switched back to normal status from alternating status). If the normal BES unit is in waiting status, the system status is displayed as SBY in the execution results of the pdls -d ha command.

3 Use the monsbystp command of HA monitor or the hatesbystp command of Hitachi HA Toolkit Extension to release the waiting status of the alternate portion. Specify in the hatesbystp command the alias of the normal BES unit that corresponds to the alternate portion to be released from standby status.

4 Use the monsbystp command of HA monitor or the hatesbystp command of Hitachi HA Toolkit Extension to release the waiting status of the normal BES unit. Specify in the hatesbystp command the unit identifier of the normal BES unit that releases the waiting status.

Examples of terminating an alternate BES unit or normal BES unit are provided below.

Example 1: Terminating a normal BES unit (normal operation, one-way alternating configuration)
[Figure]
To start the normal BES unit (UNT1) that has stopped:
  1. Use the pdstart -q command to start the normal BES unit (UNT1).
  2. Use the pdstart -q -c command to place the alternate portion of BES1 in waiting status.
Example 2: Terminating an alternate BES unit (normal operation, one-way alternating configuration)
[Figure]
To start the alternate BES unit (UNT2) that has stopped:
  1. Use the pdstart -q command to start the alternate BES unit (UNT2). The alternate portion of BES1 is also placed in waiting status.
Example 3: Terminating a unit (normal operation, mutual alternating configuration)
[Figure]
To start a unit (UNT1) that has stopped:
  1. Use the pdstart -q command to start the unit (UNT1). The alternate portion of BES2 is also placed in waiting status.
  2. Use the pdstart -q -c command to place the alternate portion of BES1 in waiting status.
Example 4: Terminating a unit (alternating, one-way alternating configuration)
[Figure]
To return to normal status:
  1. Reactivate the package for BES1. This step is necessary only when using Hitachi HA Toolkit Extension.
  2. Use the pdstart -q command to start the alternate BES unit (UNT2).
  3. Use the pdstart -q command to start the normal BES unit (UNT1).
Example 5: Terminating a unit (alternating, mutual alternating configuration)
[Figure]
* The alternate portion of BES2 is inactive.
To return to normal status:
  1. Reactivate the package for BES1. This step is necessary only when using Hitachi HA Toolkit Extension.
  2. Use the pdstart -q command to start the unit (UNT2).
  3. Use the pdstart -q command to start the unit (UNT1).
Example 6: Terminating the alternate portion (alternating, one-way alternating configuration)
[Figure]
To return to normal status:
  1. Reactivate the package for BES1. This step is necessary only when using Hitachi HA Toolkit Extension.
  2. Use the pdstart -q command to start the unit (UNT1).
Example 7: Releasing the waiting status of the alternate portion (normal operation, mutual alternating configuration)
[Figure]
To place the alternate portion of BES2, which has stopped, in waiting status:
  1. Use the pdstart -q -c command to place the alternate portion of BES2 in waiting status.
Example 8: Stopping a unit in waiting status (alternating, mutual alternating configuration)
[Figure]
* The alternate portion of BES2 is inactive.
To place a stopped unit (UNT1) in waiting status:
  1. Use the pdstart -q command to place the unit (UNT1) in waiting status.

(3) Standby-less system switchover (effects distributed) facility (terminating both the running system and the standby system)

This section explains how to terminate both the running system and the standby system.

(a) Stopping the entire system

Table 25-34 shows how the entire system is stopped when the standby-less system switchover (effects distributed) facility is used.

Table 25-34 Stopping the entire system when the standby-less system switchover (effects distributed) facility is used

Input locationCommandOptionConditionOperation
-fForcibly/abnormally terminated server?
Unit where system manager is definedpdstopNoYes1Error (a new message, such as KFPS05063-E or its equivalent, is output)
NoNo2Stops the system.
YesYesStops the system forcibly (some units are already stopped).
YesNoStops the system forcibly.
1 Even if a server has been terminated forcibly or abnormally, it is considered that there is no forcibly/abnormally terminated server if this server is in one of the following statuses:
  • Restarted at another unit and currently running.
  • Restarted at another unit and has already been stopped normally.
2 No under Forcibly/abnormally terminated server? means one of the following:
  • There are no servers in the system that have been terminated forcibly or abnormally.
  • There is a server in the system that was terminated forcibly or abnormally, but it has since been restarted at another unit and is currently running.
  • There is a server in the system that was terminated forcibly or abnormally, but it has since been restarted at another unit and has already been stopped normally.

Table 25-35 shows the processing that occurs during system termination.

Table 25-35 Processing that occurs during system termination

TargetProcessing detail
UnitWhile the system is being stopped, the system manager stops all units (by executing pdstop -u UID (-f) or its equivalent).
ServerWhen a unit is stopped while the system is being stopped, the system stops all host BESs and guest BESs at that unit (by executing pdstop -q -s server-name (-f) or its equivalent).

When the system is stopped, the accepting status for guest BESs at an accepting unit is cancelled automatically regardless of whether the system is being terminated normally terminated or forcibly. This cancellation occurs even if a guest BES is active. For this reason, you need not take any action with regard to guest BESs.

Table 25-36 shows the processing that occurs for the various back-end servers during system termination when the standby-less system switchover (effects distributed) facility is used.

Table 25-36 Processing that occurs for the various back-end servers during system termination when the standby-less system switchover (effects distributed) facility is used

Back-end server's statusProcessing
Host BESRunningStops.
Accepting statusCancels the accepting status automatically.
Guest BESRunningStops automatically.
Accepting statusCancels the accepting status automatically.

Figure 25-62 shows an example of system termination. In this example, the following servers are stopped when the system is stopped:

Additionally, the accepting status for the following servers is cancelled:

(b) Stopping a unit

Table 25-37 shows how a unit is stopped when the standby-less system switchover (effects distributed) facility is used.

Table 25-37 Stopping a unit when the standby-less system switchover (effects distributed) facility is used

Input locationCommandOptionOperation
-z-u-f
Unit where system manager is definedpdstopNoYesNoStops the target unit normally.
YesStops the target unit forcibly.
Target unitpdstopYesNoNoStops the target unit forcibly.

To stop a unit, stop all host BESs and guest BESs at the unit (by executing pdstop -q -s server-name (-f) or its equivalent).

Table 25-38 shows whether a unit can be stopped normally depending on the status of servers in the unit. When the standby-less system switchover (effects distributed) facility is used, a unit can be stopped normally regardless of whether any of its servers, host BESs or guest BESs, have stopped abnormally by themselves or have been stopped forcibly.

Table 25-38 Whether a unit can be stopped normally depending on the status of servers in the unit

Server status (both BESs and guest BESs)Can the unit be stopped normally?
Starting/stopping1On standby2Stopped3Standby-less system switchover (effects distributed) facility
NoNoNoYes
YesYes
YesNoYes
YesYes
YesNoNoNo
YesNo
YesNoNo
YesNo
1 Includes starting normally, restarting, stopping normally, stopping according to plan, being stopped forcibly, or being stopped abnormally.
2 Being blocked by mon_standby() or mon_connect().
3 Includes stopped normally, stopped according to plan, stopped forcibly, stopped abnormally, or guest area that has become unallocated after guest BES was stopped.

Table 25-39 shows the processing for the various back-end servers during unit termination when the standby-less system switchover (effects distributed) facility is used.

Table 25-39 Processing that occurs for the various back-end servers during unit termination when the standby-less system switchover (effects distributed) facility is used

Running locationBack-end server stateProcessing
Unit being stoppedRunningStops.
Accepting statusCancels the accepting status.
Other unitRunningNo change.
Accepting statusCancels the accepting status automatically.
Example 1: Stopping a unit during normal operation
Figure 25-63 shows an example of stopping a unit during normal operation.

Figure 25-63 Example of stopping a unit during normal operation

[Figure]
When a unit that has not accepted any guest BES is stopped, the following servers are also stopped:
  • Host BESs in the unit
In addition, the accepting status is cancelled for the following servers:
  • Guest BESs in the unit that are in accepting status (cancellation of accepting status)
  • Guest BESs in another unit that corresponds to the host BES in the unit being stopped (automatic cancellation of accepting status)
Example 2: Stopping a unit that has accepted a guest BES
Figure 25-64 shows an example of stopping a unit that has accepted a guest BES.

Figure 25-64 Example of stopping a unit that has accepted a guest BES

[Figure]
When a unit that has accepted a guest BES is stopped, the following servers are also stopped:
  • Host BES in the unit
  • Guest BES that is running in the unit (Automatic stop 1)
  • Host BES in another unit that corresponds to the guest BES that is running in the unit that is being stopped (Automatic stop 2)
In addition, the accepting status is cancelled for the following servers:
  • Guest BESs in the unit that are in accepting status (cancellation of the accepting status)
  • Guest BES in another unit that corresponds to the host BES in the unit being stopped (automatic cancellation of accepting status)
  • Guest BES in another unit that corresponds to the guest BES that is running in the unit being stopped (automatic cancellation of accepting status)
Example 3: Stopping a unit that has only a guest BES
Figure 25-65 shows an example of stopping a unit that has only a guest BES.

Figure 25-65 Example of stopping a unit that has only a guest BES

[Figure]
When a unit in which only a guest BES is running is stopped, the following servers are also stopped:
  • Host BES in the unit
  • Guest BES in the unit that is running (Automatic stop 1)
  • Host BES in another unit that corresponds to the guest BES that is running in the unit that is being stopped (Automatic stop 2)
In addition, the accepting status is cancelled for the following servers:
  • Guest BESs in the unit that are in accepting status (cancellation of accepting status)
  • Guest BES in another unit that corresponds to the guest BES that is running in the unit being stopped (automatic cancellation of accepting status)
(c) Stopping a server

Table 25-40 shows how a server is stopped when the standby-less system switchover (effects distributed) facility is used.

Table 25-40 Stopping a server when the standby-less system switchover (effects distributed) facility is used

Input locationCommandOptionOperation
-z-u-s-f
Unit where system manager is definedpdstopNoNoYesNoStops the target server2
YesTerminates forcibly the target servers in all active units in the HA group1.
YesYesNoStops the target server2, 3.
YesStops the target server forcibly2.
Target unitpdstopYesNoYesNoStops the target server (host BES) forcibly4.
1 Of all the active units in the HA group, servers in the running units only are stopped; the accepting status is cancelled for other units.
2 If pdstop -s(f) is used to stop a running server, the accepting status is cancelled automatically for all active units in the HA group. For Hitachi HA Toolkit Extension, ndm and rdm jointly release the standby system (accepting status). For HA monitor, HiRDB and HA monitor both cancel the accepting status.
3 Table 25-41 shows the results of server termination depending on the server status.
4 For Hitachi HA Toolkit Extension, even when pdstop -z -s is used to stop the running server, the accepting status for the server is not cancelled automatically at other units in the HA group. To cancel the accepting status, enter Hitachi HA Toolkit Extension's standby stop command (hatesbystp) at all units in the HA group.

Table 25-41 Server termination results depending on the server status

Server statusStart result
Waiting for the running system to startCancels the wait for the running system to start.
Accepting statusCancels the accepting status.
ActiveStops the server.

Table 25-42 shows the processing for the various back-end servers during server termination when the standby-less system switchover (effects distributed) facility is used.

Table 25-42 Processing that occurs for the various back-end servers during server termination when the standby-less system switchover (effects distributed) facility is used

Running locationBack-end server statusProcessing
Active unitOperation targetStops
Other unitsAccepting statusCancels the accepting status.
Example 1: Stopping a host BES
Figure 25-66 shows an example of stopping a host BES.

Figure 25-66 Example of stopping a host BES

[Figure]
When a host BES is stopped, the following server is also stopped:
  • Host BES
In addition, the accepting status is cancelled for the following servers:
  • Guest BES in another unit that corresponds to the host BES in the unit being stopped (cancellation of accepting status)
Example 2: Stopping a guest BES
Figure 25-67 shows an example of stopping a guest BES.

Figure 25-67 Example of stopping a guest BES

[Figure]
When a guest BES is stopped, the following servers are also stopped:
  • Running guest BES
  • Host BES in another unit that corresponds to the running guest BES that is being stopped (automatic stop)
In addition, the accepting status is cancelled for the following servers:
  • Guest BES in another unit that corresponds to the running guest BES that is being stopped (cancellation of accepting status)

(4) Standby-less system switchover (effects distributed) facility (terminating only the standby system)

This section explains how to stop only the standby system.

As when using the standby system switchover facility or standby-less system switchover (1:1) facility, the monsbystp command of HA monitor can be used to stop the standby system. When the standby-less system switchover (effects distributed) facility is used, you can also perform the operation from the unit where the system manager is defined.

Table 25-43 shows how to terminate the standby system when the standby-less system switchover (effects distributed) facility is used.

Table 25-43 Terminating the standby system when the standby-less system switchover (effects distributed) facility is used

Input locationCommandOperation targetOperation
Host where the operation-target server is locatedmonsbystp*Back-end server in accepting statusCancels the accepting status for a guest BES
Unit where system manager is definedpdstop -u -s
* When you use Hitachi HA Toolkit Extension, use the hatesbystp command.
Example 1: Example of cancelling the accepting status for a guest BES

Figure 25-68 shows an example of cancelling the accepting status for a guest BES.

Figure 25-68 Example of cancelling the accepting status for a guest BES

[Figure]

To cancel the accepting status for a guest BES, enter HA monitor's monsbystp command. When you use Hitachi HA Toolkit Extension, enter the hatesbystp command.

Example 2: Example of stopping a host BES of the standby system

Figure 25-69 shows an example of stopping a host BES of the standby system.

Figure 25-69 Example of stopping a host BES of the standby system

[Figure]

To stop a host BES of the standby system, enter HA monitor's monsbystp command. When you use Hitachi HA Toolkit Extension, enter the hatesbystp command.