Unfamiliar recovery message afer server crash

[Arthur Ward]

For unknown reasons, our PG server died overnight. (PG 7.4.2, RedHat 8,
ext3 ordered data mode, and battery backed RAID5) When it came back up, I
saw an unfamiliar line in the recovery output from Postgres:

Apr 20 11:28:14 postgres: [6203] LOG: database system was interrupted at
2004-04-20 05:50:59 CDT
Apr 20 11:28:14 postgres: [6203] LOG: checkpoint record is at 7C/CF55FB6C
Apr 20 11:28:14 postgres: [6203] LOG: redo record is at 7C/CF55FB6C; undo
record is at 0/0; shutdown FALSE
Apr 20 11:28:14 postgres: [6203] LOG: next transaction ID: 9374951; next
OID: 158516389
Apr 20 11:28:14 postgres: [6203] LOG: database system was not properly
shut down; automatic recovery in progress
Apr 20 11:28:14 postgres: [6203] LOG: redo starts at 7C/CF55FBAC
Apr 20 11:28:17 postgres: [6203] LOG: unexpected pageaddr 7C/C7A0A000 in
log file 124, segment 207, offset 10526720
Apr 20 11:28:17 postgres: [6203] LOG: redo done at 7C/CFA09F7C
Apr 20 11:28:20 postgres: [6203] LOG: database system is ready

What's the significance of the "unexpected pageaddr" item?


---------------------------(end of broadcast)---------------------------
TIP 1: subscribe and unsubscribe commands go to ***@postgresql.org

[Tom Lane]

Post by Arthur Ward
Apr 20 11:28:14 postgres: [6203] LOG: database system was not properly
shut down; automatic recovery in progress
Apr 20 11:28:14 postgres: [6203] LOG: redo starts at 7C/CF55FBAC
Apr 20 11:28:17 postgres: [6203] LOG: unexpected pageaddr 7C/C7A0A000 in
log file 124, segment 207, offset 10526720
Apr 20 11:28:17 postgres: [6203] LOG: redo done at 7C/CFA09F7C
Apr 20 11:28:20 postgres: [6203] LOG: database system is ready
What's the significance of the "unexpected pageaddr" item?

Just that the first indication of end-of-XLOG was finding a stale page
header in a recycled XLOG file. The above looks fine to me. (The XLOG
recovery code tends to err on the side of verboseness... there are a lot
of messages it can print that might boggle a novice but are not really
indicative of problems.)

regards, tom lane

---------------------------(end of broadcast)---------------------------
TIP 3: if posting/reading through Usenet, please send an appropriate
subscribe-nomail command to ***@postgresql.org so that your
message can get through to the mailing list cleanly