search

OPMDG / check_pgactivity

Nagios remote agent
http://opm.readthedocs.io/probes/check_pgactivity.html
PostgreSQL License
175 stars 51 forks source link

hot_standby_delta with log shipping #284

Open Krysztophe opened 3 years ago

Krysztophe commented 3 years ago

On a replica with log shipping only, the ''hot_standby_delta'' service gives strange results:

receive delta service:nagios_defo13_2 | 656.00 MiB | 100.00 MiB | 1024.00 MiB
replay delta service:nagios_defo13_2 | 0.00 B | 200.00 MiB | 2.00 GiB

In fact , for log shipping, pg_last_wal_receive_lsn is stuck at the end of the recovery.

#  SELECT pg_last_wal_receive_lsn(),  pg_last_wal_replay_lsn() ;
 pg_last_wal_receive_lsn | pg_last_wal_replay_lsn 
-------------------------+------------------------
 1C/EB0001C0             | 1D/11000000

It seems that replacing ''coalesce''' by ''greatest'' in the query works (see https://github.com/OPMDG/check_pgactivity/blob/8009e0e6ab209578edde25f8bd58b80fc2a06955/check_pgactivity#L4227 )

Option: get rid of the "receive" perfdata when pg_last_wal_receive_lsn < pg_last_wal_replay_lsn

Krysztophe commented 3 years ago

BTW this cas occurs only after a switch from streaming to log shipping without restarting the replica.