transaction-wraparound data loss
transaction-wraparound data loss
От:
"Brian E. Pangburn" <bpangburn@nqadmin.com>
Дата:
I run the following nightly using cron against my 7.3.3 databases: vacuumdb -a -f -z The other night, I got the following (db names have been changed to protect the innocent): Vacuuming db1 WARNING: Some databases have not been vacuumed in over 2 billion transactions. You may have already suffered transaction-wraparound data loss. VACUUM Vacuuming db2 VACUUM Vacuuming db3 VACUUM Vacuuming template1 VACUUM I run this nightly and I don't think there is any way that I could have processed 2 billion transactions in a single day. The first database "db1" really isn't even used actively. Could the error message have been generated in error? Is there something I should check or do? Does this indicate that I'm going to have catastrophic errors with one or all of my databases? Any insights are much appreciated. Brian E. Pangburn
Re: transaction-wraparound data loss
От:
Tom Lane <tgl@sss.pgh.pa.us>
Дата:
"Brian E. Pangburn" writes: > I run the following nightly using cron against my 7.3.3 databases: > vacuumdb -a -f -z > The other night, I got the following (db names have been changed to protect > the innocent): > WARNING: Some databases have not been vacuumed in over 2 billion > transactions. Hm, that should not happen if you've been doing routine vacuumdb -a. Could we see the contents of pg_database? regards, tom lane