Yes, you do recall correctly. It is Solaris 10 and Postgres 8.2.3.
I'm going to run a pg_dumpall and keep my fingers crossed. I'm getting similar errors when I try to do other things like vacuumdb, so maybe pg_dumpall won't run either.
Carol
On Sep 23, 2008, at 6:13 PM, Tena Sakai wrote:
Hi Carol,
I don't have answer...
I have had a similar problem and whatever I did didn't work. As I recall, the files in pg_clog directory started from 0000 and the number got bigger by 1 (hex) and there was a gap between the most recent one and the second recent one. (Is that how you know that you are missing about 30 files?) I suspect the reason why it didn't work for me was because the last file was incomplete. I wound up upgrading to the latest postgres and restored from a backup file I made by pg_dumpall. My platform is linux (do I recall yours is solaris?) and the postgres version I was running was 8.3.0.
The lesson I learned was to keep an eye on the pg_clog directory. Since the installation of 8.3.3 there has been only one file (0000), though the size keeps growing (I think by 8k bytes).
