View Issue Details
| ID | Project | Category | View Status | Date Submitted | Last Update |
|---|---|---|---|---|---|
| 0000816 | Pgpool-II | General | public | 2023-10-25 03:41 | 2023-11-02 04:14 |
| Reporter | henri_pilon@belll.net | Assigned To | |||
| Priority | normal | Severity | minor | Reproducibility | always |
| Status | new | Resolution | open | ||
| Product Version | 4.3.7 | ||||
| Summary | 0000816: pgpool-II 4.3 relationship with postgresql-12 and archive_cleanup_command off the primary PostgreSQL system | ||||
| Description | My apologies if this question has been asked before , and looking for some guidance as I think this is more postgresql 12 issue maybe. Configuration - 3 nodes cluster (primary and 2 standby systems) with pgpool-II 4.3 and PostgreSQL 12 postgresql.conf off each systems. shared_buffers = 1024MB work_mem = 50MB maintenance_work_mem = 256MB effective_cache_size = 4096MB max_wal_size = 2048MB max_connections = 1000 checkpoint_timeout = 20min checkpoint_completion_target = 0.9 autovacuum_vacuum_threshold = 5000 autovacuum_analyze_threshold = 5000 autovacuum_analyze_scale_factor = 0.2 log_min_duration_statement = 3000 commit_delay = 500 max_locks_per_transaction = 512 password_encryption = md5 wal_level = replica archive_mode = on archive_command = 'cp %p /var/lib/pgsql/archivedir/%f' hot_standby = on max_wal_senders = 10 max_replication_slots = 10 max_wal_size=200 wal_keep_segments = 10 port = 5432 wal_log_hints = on restore_command = 'cp /var/lib/pgsql/archivedir/%f %p' archive_cleanup_command = 'pg_archivecleanup /var/lib/pgsql/archivedir %r' include_if_exists = '/var/lib/pgsql/12/data/myrecovery.conf' I understand the pgpool-II myrecovery.conf overwriting the restore_command off each standby systems. physical slot replication is being utilize across the 2 standby systems, letting pgpool-II perform this configuration with the appropriate pcp command Our issue is that off the primary server the dedicate disk reserve for postgres keeps filling up, especially the /var/lib/pgsql/archivedir folder. My understanding the archive_cleanup_command action is only perform off the standby systems. The way we are currently are managing the issue is by periodically performing a pg_basebackup off the primary system to a backup folder and then performing the archive_cleanup_command manually looking for the .backup file in the archivedir folder. We can see the du -h /var/lib/pgsql/12/data/pg_wal file in this folder get deleted (purge) when the folder size gets close to the max_wal_size size, but du -h /var/lib/pgsql/archivedir keep growing. I did several google search and unable to find any relevant information around this subject in relation to pgpool-ii. What’s the proper way of managing the archivedir space off the primary system? What’s the best practice for this area and any guidance would be appreciated ? Hope my question make sense ... | ||||
| Steps To Reproduce | using pg_bench, to create and re-create transaction in PostgreSQL keep monitoring disk space off the primary PostgreSQL system. du -h /var/lib/pgsql/archivedir vs du -h /var/lib/pgsql/12/data/pg_wal | ||||
| Tags | No tags attached. | ||||
|
|
I think at this point I'm confident the "achivedir" folder space needs to be managed outside of both PGPOOL-II and PostgreSQL itslef. |
| Date Modified | Username | Field | Change |
|---|---|---|---|
| 2023-10-25 03:41 | henri_pilon@belll.net | New Issue | |
| 2023-11-02 04:14 | henri_pilon@belll.net | Note Added: 0004441 |
