Hi,
We have a PG 8.4 database approx 5TB in size.
We were recently testing our restore procedure against our latest dump. The dumps are taken using the Continuous
Archivingmethod with base dumps taken using tar. Our tar script is set up to ignore missing/modified files but should
stopon all other errors.
We are testing the restore on a completely separate machine of a similar spec to the server on which the dumps are
made.
On our first attempt we were using PG 8.4.0. We struck problems on restore with the error 'Unexpected timeline ID 0 in
logfile'. After some googling we identified that this was a known bug in PG 8.4.0 so we upgraded the target machine to
8.4.7and began the recovery again.
On the second attempt the recovery went through smoothly. However, after starting the application again we started to
getthe following error:
ERROR: missing chunk number 0 for toast value 382548694 in pg_toast_847386
The table it is trying to access is one in which our application stores various schema information that is used each
timethe application starts up.
As far as we know the base dump is good and the DB did not complain about any of the WAL files. We are now nervous that
ourbackups are no good, so we want to get to the bottom of it. What can be the reason that these backups are not
restoringproperly?
Thanks in advance,
Regards
Mike Harris
