Node restart causes application - ERROR: restart transaction:

Hello,

When we restart CDB nodes ( 1 node at a time), the application sees following error. We have ELB in front of the cluster which app uses to connect. ( Timeout - 5 seconds, Interval|- 10 seconds, Unhealthy threshold - 2). Also tested different value for “server.shutdown.drain_wait” (60s, 1m30s). But we are still getting the error.

Caused by: org.postgresql.util.PSQLException: ERROR: restart transaction: TransactionRetryWithProtoRefreshError: TransactionAbortedError(ABORT_REASON_TIMESTAMP_CACHE_REJECTED_POSSIBLE_REPLAY): “sql txn” id=eb6f6052 key=/Table/94/1/"dddd/“O”/0 rw=true pri=0.02494864 stat=PENDING epo=0 ts=1565812919.093580066,1 orig=1565812919.090634522,0 max=1565812919.093580066,0 wto=false seq=3
at org.postgresql.core.v3.QueryExecutorImpl.receiveErrorResponse(QueryExecutorImpl.java:2440)

Hey @mshah,

Looks like this error is related to an unresolved transaction.

It’s hard to debug without seeing what types of sql operations your app is doing.

I suggest using some client side intervention.

What version of CRDB are you using?

Thanks,
Matt

CRDB is v19.1.3.

Will read through the link you provided.

Sounds good, let me know if it resolves.

Thanks,
Matt