Integrated replicat hangs sporadically (Doc ID 2347717.1)

Last updated on JANUARY 30, 2023

Applies to:

Symptoms

Time after time some of the integrated replicat processes completely hangs with the following symptoms

1) Writing to it's *.rpt log suddenly stops without any error or message.

2) Trying to communicate with replicat process by issuing SEND command results in a "Timeout waiting for message" :

info <replciat_name>

REPLICAT <replicat_name> Last Started 2015-09-04 01:17 Status RUNNING
INTEGRATED
Checkpoint Lag 00:00:03 (updated 00:29:27 ago)
Process ID 9926
Log Read Checkpoint File <CHECKPOINT_FILE>
2015-09-14 08:00:33.434204 RBA 863949562

send <replicat_name>, stats

Sending STATS request to REPLICAT <replicat_name> ...
ERROR: sending message to REPLICAT <replicat_name> (Timeout waiting for message).

3) No errors are reported either in ggserr.log or in alert.log.

4) Killing and restarting replicat process resolves the issue for a while.

5) If we check STATE column of some v$gg_apply_* we can see that almost all of them have value "IDLE", except of v$gg_apply_receiver view. It's column STATE has value "Waiting for memory".

6) There was enough free memory in the Steams Pool at the moment of hanging:

select * from v$sgastat where pool like 'streams pool' and name='free memory'

streams pool free memory 12849013848 0

Cause

My Oracle Support provides customers with access to over a million knowledge articles and a vibrant support community of peers and Oracle experts.