Robert Haas <robertmh...@gmail.com> writes: > Yeah, this is a bug. It seems that the WAL summarizer process, when > restarted, wants to restart from wherever it was previously > summarizing WAL, which is correct if that WAL is still around, but if > summarize_wal has been turned off in the meanwhile, it might not be > correct. Here's a patch to fix that.
This comment seems to be truncated: + /* + * If we're the WAL summarizer, we always want to store the values we + * just computed into shared memory, because those are the values we're + * going to use to drive our operation, and so they are the authoritative + * values. Otherwise, we only store values into shared memory if they are + */ + LWLockAcquire(WALSummarizerLock, LW_EXCLUSIVE); + if (am_wal_summarizer|| !WalSummarizerCtl->initialized) + { regards, tom lane