We have a Postgres 8.4 installation which hosts both our Defect tracking and Code management (Surround) systems. We are looking at converting and migrating Surround to MSSQL on another server. Our Surround structure contains multiple DBs one of which is 121GB, which in order to carry out the migration needs to be purged of its cache.
I have carried out a restore onto my local machine, utilising pg_dump, which restores the DB to a size of 82GB. The premise here was to see if carrying out this process would be efficient, before the migration / conversion, which it very much is.
The problem I'm having is I can't carry out the full VACUUM against this DB, due to the HD space considerations with regards to duplicating its size for this process.
I have come across a very helpful post below, where I can re-initialize the data dictionary after a pg_dumpall and then restore, removing the need for extra HD space required for the VACUUM. The problem here however is that the other DBs will also be included in this process, which is too much downtime. Is there something similar I can carry out against just the effected DB to purge its cache?
I need to run VACUUM FULL with no available disk space
Thanks for your time in advanced.
