Nonstop Database, HiRDB Version 9 Installation and Design Guide
The organization of processes started by HiRDB is shown below.
The processes started by a HiRDB parallel server configuration are listed in the following table.
Table B-5 Processes started by a HiRDB parallel server configuration (system server) (1/2)
| Process name | Description | Number of processes | Server name | Does process start for each server? | ||
|---|---|---|---|---|---|---|
| Notation in manual | Process name | |||||
| MGR | Non-MGR | |||||
| Process server process | pdprcd | Manages HiRDB related processes | 1 | _prc | Y | Y |
| Post-processing process | pdrsvre | Performs cleanup processing after abnormal termination of a process | If fixed is specified in pd_process_terminator: value of pd_process_terminator_max If resident is specified in pd_process_terminator: from 3 to the number of processes defined by HiRDB If nonresident is specified in pd_process_terminator: from 0 to the number of processes that terminated abnormally |
_admrsvr | Y | Y |
| HiRDB start process for server mode system switchover 1 | pdstart2d | Controls start of HiRDB processes linked to cluster software
|
1 (none when online) | _pdstrt2 | Y | Y |
| HiRDB start process for server mode system switchover 2 | pdstart2a | Controls start of HiRDB processes linked to cluster software
|
1 (none when online) | _pdst2a1 | N | Y |
| HiRDB start process for server mode system switchover 3 | pdsvstartd | Controls start of HiRDB processes linked to cluster software
|
If activeunits is specified in pd_ha_agent, 1; otherwise, 0 | _pdsvstd | N | Y |
| XDS log output process#14 | pdprctee | Controls output of XDS log | If the pdxds operand is specified for a BES within a unit, 1; otherwise, 0 | -- | Y | Y |
| Command execution process | pdcmdd | Controls execution of update log reflection command | If the pdxds operand is specified for a BES within a unit, 1; otherwise, 0 | _cmdd | Y | Y |
| Cluster synchronization monitoring process | pdxcl | Synchronization monitoring of running XDS and standby XDS | If the pdxds operand is specified for a BES within a unit, 1; otherwise, 0 | _xcl | Y | Y |
| Troubleshooting information acquisition process | pdprfd | Controls the troubleshooting function | 1 | -- | Y | Y |
| Message log server process | pdmlgd | Controls message output (starts when local is specified for pd_mlg_msg_log_unit) |
1 | _mlg | Y | Y |
| System manager (MGR) process | pdrdmd | Controls unit start/stop and manages connected users (This might also be notated as a name server or a node manager.) |
1 | _rdm | Y | N |
| Node manager (non-MGR unit) process | pdndmd | Controls unit start/stop and manages connected users (This might also be notated as a name server.) |
1 | _ndm | N | Y |
| Status server process | pdstsd | Controls I/O of status files for units | 1 | _sts0 | Y | Y |
| Scheduler process | pdscdd | Allocates back-end server, dictionary server, and front-end server processes (This might also be notated as a lock server.) |
1 | _scd | Y | Y |
| Transaction server process | pdtrnd | Controls transactions | 1 | _trn | Y | Y |
| Transaction recovery process | pdtrnrvd | Controls transaction completion/recovery | For an FES: from 1 to the number of crashed pdfes#1 For a DS: from 1 to the number of crashed pddic#1 For a BES: from 1 to the number of crashed pdbes#1 |
_trnrcv | Y | Y |
| Audit trail management server process | pdaudd | Manages audit trails | If pd_aud_file_name is specified, 1; otherwise, 0 | _aud _auz#2 |
Y | Y |
| Audit trail automatic load process | pdaudld | Controls start of pdload for automatic data load | If Y is specified for pd_aud_auto_loading in keeping with the conditions of the audit trail management server process, 1; when N or nothing is specified, 0 | _audld | Y | N |
| Unit monitoring process | pdrdma | Monitors whether the HiRDB unit is running | 1 (when the number of units is 1, 0) | _rdmck | Y | N |
Table B-6 Processes started by a HiRDB parallel server configuration (system server) (2/2)
| Process name | Description | Number of processes | Server name | Does process start for each server? | |||
|---|---|---|---|---|---|---|---|
| Notation in manual | Process name | ||||||
| BES | DS | FES | |||||
| Log server process | pdlogd | Controls system log acquisition and log-related processes | 1 | _logN _lozN#2 |
Y | Y | Y |
| Deferred write process | pd_buf_dfw | Background write to DB storage disk | 1 | 1dfwN | Y | Y | N |
| Asynchronous READ process | pd_ios_ard | Asynchronous READ facility | Value of pd_max_ard_process | 1ardN | Y | Y | N |
| Parallel WRITE process for deferred write processing | pd_buf_awt | Facility for parallel writes in deferred write processing | Value of pd_dfw_awt_process | 1awtN | Y | Y | N |
| REDO process | pd_rcv_rd | DB roll forward at full rerun | MIN(number of connected disks, value of pd_max_recover_process) Number of connected disks: Number of character special files that define RDAREAs If the value of pd_rdarea_open_attribute_use is Y, the value of pd_max_recover_process#3 |
2rrnM | Y | Y | N |
| Log swap process | pdlogswd | Manages allocation, release and I/O of system log-related files and acquires syncpoint dump | 1 | _logsN _lozsN#2 |
Y | Y | Y |
| Deadlock monitoring process | pdlckmnd | Detects deadlock when lock processing is distributed | For a FES: If 2 or more is specified for pd_fes_lck_pool_partition and if Y is specified for pd_lck_deadlock_check, 1; otherwise, 0 For a DS or BES: If 2 or more is specified for pd_lck_pool_partition and if Y is specified for pd_lck_deadlock_check 1; otherwise, 0 |
_lckmnN _lckmzN#2 |
Y | Y | Y |
Table B-7 Processes started by a HiRDB parallel server configuration (user server)
| Process name | Explanation | Number of processes | Server name | Does process start for each server? | |||
|---|---|---|---|---|---|---|---|
| Notation in manual | Process name | ||||||
| BES | DS | FES | |||||
| Back-end server process | pdbes | Access to database | MAX (value of pd_max_users, value of pd_max_bes_process) | Server name#4 | Y | N | N |
| Dictionary server process | pddic | Batch management of dictionary table | MAX (value of pd_max_users, value of pd_max_dic_process) | Server name#4 | N | Y | N |
| Front-end server process | pdfes | SQL processing, instructions to back-end server | Value of pd_max_users | Server name#4 | N | N | Y |
Table B-8 Processes started by a HiRDB parallel server configuration (XDS server)
| Process name | Explanation | Number of processes | Server name | Does process start for each server? | |||
|---|---|---|---|---|---|---|---|
| Notation in manual | Process name | ||||||
| BES | DS | FES | |||||
| XDS process | pdxds | SQL processing, manages memory database | For each pdxds operand specification, 1#12 | Server name#13 | Y | N | N |
Table B-9 Processes started by a HiRDB parallel server configuration (utility server)
| Process name | Explanation | Number of processes | Server name | Does process start for each server? | |||
|---|---|---|---|---|---|---|---|
| Notation in manual | Process name | ||||||
| BES | DS | FES | |||||
| pdinit control process | pdinitd | Initialization utility execution process | 1 | 0minit0 | N | Y | N |
| pdinit execution process | pdinitb | Initialization utility BES-side execution process | 1 to 2 | 0sinit0 | Y | N | N |
| pdcopy backup output process | pdcopyb | Backup file output (starts at the backup output destination specified in pdcopy (-b option)) |
Degree of duplication |
0bcpy?0#5 | N | N | N |
| pdcopy database read process | pdcopyr | Database reads | Number of servers subject to copy#6 | 0rcopyN | Y | Y | N |
| pdrstr backup read process | pdrstrb | Backup file reads (starts at the backup read destination specified in pdrstr (-b option)) |
Number of concurrent pdrstr command executions | 0brstr0 | N | N | N |
| pdrstr master directory RDAREA read process | pdrstrm | Reads master directory RDAREA#7 | Number of concurrent pdrstr command executions | 0mrstr0 | N | Y | N |
| pdrstr unload log read process | pdrstrl | Reads unload log file#8 | Number of concurrent pdrstr command executions | 0lrstr0 | N | N | N |
| pdrstr database write process | pdrstrr | Database writes | Number of servers subject to recovery#9 | 0brstrN | Y | Y | N |
| pdrstr master directory RDAREA write process | pdrstrw | Writes to master directory RDAREA#10 | 1 | 0wrstr0 | N | Y | N |
| pdload control process | pdloadm | Controls data loads | Number of concurrent pdload command executions | 0mload0 | Y | Y | Y |
| pdrorg control process | pdrorgm | Controls database reorganization (unloading, reloading, index reorganization/re-creation, free page release, and making global buffers resident) | Number of concurrent pdrorg, pdreclaim, and pdpgbfon command executions | 0mrorg0 | Y | Y | Y |
| pdrbal control process | pdrbalm | Controls rebalancing | Number of concurrent pdrbal command executions | 0mrbal0 | N | Y | N |
| pdgetcst control process | pdgcstm | Collects optimization information | Number of concurrent pdgetcst command executions | 0mgcst0 | N | Y | N |
| pddbst control process | pddbst1 | Controls database condition analysis (starts at MGR unit) |
Number of concurrent pddbst command executions | 0mdbst0 | N | N | N |
| pdexp control process | pdexpm | Controls dictionary export/import (starts in unit that has export file) |
1 | 0mexp0 | N | N | N |
| pdplgexe control process | pdplgexm | Plug-in utility execution control | Number of concurrent executions of commands provided by plug-ins | 0mplge0 | N | Y | N |
| pdorend control process | pdorendm | Online reorganization Reflection control |
1 | 0more0 | Y | N | N |
| pdorend application process | pdorendl | Online reorganization Reflection processing |
Degrees of duplication (number of specifications of the pdorend command's -m option) | 0lore?0#11 | Y | N | N |
| pdutlcex control process | pdutlexm | Controls utility execution | Number of parallel executions (pd_import_export_parallel operand value of XDS database definition) |
0mucex | N | N | N |
All Rights Reserved. Copyright (C) 2012, 2015, Hitachi, Ltd.