Can you reproduce that one with 7fcbf6a^ (i.e before xlogreader got
split off?).
Yes, it is reproducible before the xlog reader split. Just an additional report, the master jumps correctly to the new timeline.
> The replication delays are still here.
That one is caused by this nice bug, courtesy of yours truly: diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c index 90ba32e..1174493 100644 --- a/src/backend/access/transam/xlog.c +++ b/src/backend/access/transam/xlog.c @@ -8874,7 +8874,7 @@ retry: /* See if we need to retrieve more data */ if (readFile < 0 || (readSource == XLOG_FROM_STREAM && - receivedUpto <= targetPagePtr + reqLen)) + receivedUpto < targetPagePtr + reqLen)) { if (StandbyMode) {
I didn't notice because I had a testscript inserting stuff continuously and it cause at most lagging by one record...