Hi,
Here are some improvements we can make to pg_receivewal that were
emanated after working with it in production environments:
1) As a user, I, sometimes, want my pg_receivewal to start streaming
from the LSN that I provide as an input i.e. startpos instead of it
calculating the stream start position 1) from its target directory or
2) from its replication slot's restart_lsn or 3) after sending
IDENTIFY_SYSTEM on to the primary.
This will particularly be useful when the primary is down for some
time (for whatever reasons) and the WAL files that are required by the
pg_receivewal may have been removed by it (I know this situation is a
bit messy, but it is quite possible in production environments). Then,
the pg_receivewal will
calculate the start position from its target directory and request the
primary with it, which the primary may not have. I have to intervene
and manually delete/move the WAL files in the pg_receivewal target
directory and restart the pg_receivewal so that it can continue.
Instead, if pg_receivewal can accept a startpos as an option, it can
just go ahead and stream from the primary.
2) Currently, RECONNECT_SLEEP_TIME is 5sec - but I may want to have
more reconnect time as I know that the primary can go down at any time
for whatever reasons in production environments which can take some
time till I bring up primary and I don't want to waste compute cycles
in the node on which pg_receivewal is running and I should be able to
just set it to a higher value, say 5 min or so, after which
pg_receivewal can try to perform StreamLog(); and attempt connection
to primary.
Thoughts?
Regards,
Bharath Rupireddy.
