Thread: could not access status of transaction 1118722281
Running 9.0.2 on Centos. I just discovered this in my production error log. Starting about 45 minutes ago, I got 70 of these, within 2 seconds: 28652 2011-04-04 21:47:29 EDT [33]WARNING: PD_ALL_VISIBLE flag was incorrectly set in relation "pg_toast_49338181" page 16820 These warnings were immediately proceeded by this, which has continuously repeated every 15 seconds since then: 8895 2011-04-04 22:15:28 EDT [1]ERROR: could not access status of transaction 1118722281 8895 2011-04-04 22:15:28 EDT [2]DETAIL: Could not open file "pg_clog/042A": No such file or directory. 8895 2011-04-04 22:15:28 EDT [3]CONTEXT: automatic vacuum of table "mcore.pg_toast.pg_toast_48975830" I checked and the pg_clog files start at 04BF and run through 0A57 (1,433 files) Any help would be greatly appreciated. Gordon -- View this message in context: http://postgresql.1045698.n5.nabble.com/could-not-access-status-of-transaction-1118722281-tp4283137p4283137.html Sent from the PostgreSQL - general mailing list archive at Nabble.com.
Turns out this was most likely the pg_upgrade bug. In our case, I was able to dump and recreate the table in question. Since then, this has been made public: http://wiki.postgresql.org/wiki/20110408pg_upgrade_fix -- View this message in context: http://postgresql.1045698.n5.nabble.com/could-not-access-status-of-transaction-1118722281-tp4283137p4293709.html Sent from the PostgreSQL - general mailing list archive at Nabble.com.