Eight weeks ago we had a SAN failure and had to send our drive to a data recovery service. The data has come back. It was copied to the new SAN using the same folder structure as before. The Owner and permissions were changed on the files to match what they were previously. A variety of errors showed up when trying to access the newly recovered DB(depending on what I messed with, the errors were something like).
invalid page header in block xx of relation "pg_class",
could not open relation 16388/16390/5175819
catalog is missing 3 attributes for relid 6787355
I only need the data from one table(granted the table is huge). I tried using "Recovery for Postgresql" software, that probably would have worked great except that a critical field in the table is bytea, and I don't think the recovery software works on bytea.
I am desperatly in need of some help. I was able to get pg_dump to run today against the single table, but it repeatly died with 'unexpected chunk number 32 (expected 19) for toast value 6782772. I tired setting zero_damaged_pages = true( and zero_damaged_pages = on) but no luck.
What I need is to be able to export the data from one table (documentdata2) and import it into a new database.
Any ideas? using postgres 8.2 on a Suse Linux box.
pgsql-novice by date
|Next:||From: Lew||Date: 2010-07-27 12:32:38|
|Subject: Re: Select only active entries|
|Previous:||From: Joshua Tolley||Date: 2010-07-26 21:18:32|
|Subject: Re: Missing Functions|