Hi all,
we’re using Inter-SyncGateway Replication with several SyncGateways v3.1.2 instances.
We’ve set up push and pulls replication and sometimes seems like that they need a restart.
Even if the replications are in the Running status, seems like they lose the checkpoints and doesn’t update the “seq”. Hence, we usually stop and restart the replication to force the seq update.
What could be the problem? Is this a known issue?
This isn’t a known issue. It’s difficult to diagnose from the information provided. If you have a support entitlement, I’d suggest filing a support ticket to investigate this in detail.
Barring that, your best option is probably to review the ISGR-related stats to identify the point at which replication appears to “need a restart”, then review the SGW logs around that time to check for any unexpected warnings/errors.
Hi @adamf, thanks for the reply.
Yes, I’m moving in that way and engaged the support.
The only thing that I know at the moment is that the “committed_seq” on the databases, at a certain time stops to be aligned in the replications.
The “last_seq_pull” and the “last_seq_push” updates, respectively for pull and for push replications, are extremely slow or stopped at all.
Hence, with a restart of the replications the seq are recomputed and forced to align.