Обсуждение: Troubling On-line Backup LOG messages
After a successful same-server test of the postgres on-line backup process, I tried a test from a remote server to simulate a process I'll need to do in a production environment soon. After a step-by-step on-line backup, I saw the following in my logs: 2006-03-21 10:38:53 CST 1412 :LOG: archive recovery complete 2006-03-21 10:38:53 CST 1412 :LOG: could not truncate directory "pg_multixact/offsets": apparent wraparound 2006-03-21 10:38:53 CST 1412 :LOG: could not truncate directory "pg_multixact/members": apparent wraparound 2006-03-21 10:39:22 CST 1412 :LOG: database system is ready2006-03-21 10:39:22 CST 1412 :LOG: transaction ID wrap limit is 2147484146, limited by database "postgres" followed by dozens of these: 2006-03-21 10:39:22 CST 2499 :LOG: invalid server process ID -1 None of these are errors, but they don't seem like healthy messages nonetheless. I don't see much discussion of them in the archives. Are they cause for concern? -- Thomas F. O'Connell Database Architecture and Programming Co-Founder Sitening, LLC http://www.sitening.com/ 3004 B Poston Avenue Nashville, TN 37203-1314 615-260-0005 (cell) 615-469-5150 (office) 615-469-5151 (fax)
"Thomas F. O'Connell" <tfo@sitening.com> writes: > 2006-03-21 10:38:53 CST 1412 :LOG: archive recovery complete > 2006-03-21 10:38:53 CST 1412 :LOG: could not truncate directory > "pg_multixact/offsets": apparent wraparound > 2006-03-21 10:38:53 CST 1412 :LOG: could not truncate directory > "pg_multixact/members": apparent wraparound What file name(s) appear in those directories? Do you continue to get the same log messages during subsequent checkpoints? > followed by dozens of these: > 2006-03-21 10:39:22 CST 2499 :LOG: invalid server process ID -1 I've seen this off and on in CVS HEAD, mostly just after provoking a backend crash ;-). The message is evidently from the pgstats subsystem. I think it has to do with backends trying to send stats messages before they're fully initialized, but haven't tracked it down yet. It should be pretty harmless in any case. regards, tom lane
On Mar 21, 2006, at 11:54 AM, Tom Lane wrote: > "Thomas F. O'Connell" <tfo@sitening.com> writes: >> 2006-03-21 10:38:53 CST 1412 :LOG: archive recovery complete >> 2006-03-21 10:38:53 CST 1412 :LOG: could not truncate directory >> "pg_multixact/offsets": apparent wraparound >> 2006-03-21 10:38:53 CST 1412 :LOG: could not truncate directory >> "pg_multixact/members": apparent wraparound > > What file name(s) appear in those directories? Do you continue to get > the same log messages during subsequent checkpoints? pg_multixact/offsets/0001 pg_multixact/members/0002 Thus far, there haven't been any subsequent checkpoints. These occurred immediately after recovery of an on-line backup, and I haven't really done much with the new cluster because it just exists for proof of concept. -- Thomas F. O'Connell Database Architecture and Programming Co-Founder Sitening, LLC http://www.sitening.com/ 3004 B Poston Avenue Nashville, TN 37203-1314 615-260-0005 (cell) 615-469-5150 (office) 615-469-5151 (fax)