Backend Connection Failed

Hi Folks,

We have HA with Master/Slave database along with pgpool…where application were connecting via pgpool with VIP…

Suddenly VIP was unable to telnet and application went in waiting state…We have received frequently below error in pgpool.log file.

2021-07-17 08:12:12: pid 1618597: WARNING: write on backend 1 failed with error :“Success”
2021-07-17 08:12:12: pid 1618597: DETAIL: while trying to write data from offset: 0 wlen: 5
2021-07-17 08:12:12: pid 1618597: WARNING: write on backend 1 failed with error :“Success”
2021-07-17 08:12:12: pid 1618597: DETAIL: while trying to write data from offset: 0 wlen: 5

Could someone help here why VIP went down any causes…

1 Like

What does your postgres logs say?

I’m just trying to establish there’s nothing going on in the backend i.e. no error messages in postgres logs confirms whatever is happening has nothing to do with the database server itself.

Here’s one example, simple and to the point:
cat $PGDATA/log | grep -E 'ERROR|FATAL' | less -SN

2 Likes

While writing to a socket, sometimes write() returns 0 this could happen in the field. Before we
treat this as an error, checking the Postgres logs would be worth it. If there are no error messages in Postgres logs it is better to treat this WARNING message as normal.

1 Like