What would happen if a recycled segment gets assigned a newer timeline than
the one we are currently recovering from? In the reported erroneous
behavior, that is what happens causing XLogFileReadAnyTLI() to return such
bogus segment causing the error. Since, expectedTLIs contains a newer
timeline at its head, and if such a wrongly recycled (that is one bearing a
newer timeline than curFileTLI) segment exists in pg_xlog,
XLogFileReadAnyTLI() would return fd of that segment itself causing the
error in question. In next retry, the same thing would happen and prevent
standby startup from proceeding any further.
So, should we also look at the segment recycling code to see how it names
the recycled segments (what timelineID does it use) ? For example, in
CreateRestartPoint() where segments are recycled?
--
View this message in context:
http://postgresql.1045698.n5.nabble.com/Failing-start-up-archive-recovery-at-Standby-mode-in-PG9-2-4-tp5753110p5753358.html
Sent from the PostgreSQL - hackers mailing list archive at Nabble.com.