I discovered this issue a bit more. -j option is slowing down BLOBs
restoring. It's about 1000x times slower if you specify this option.
Does anybody plan to fix it?
> I have encountered a problem while restoring the database. There is a
> table that contains XML data (BLOB), ~ 3 000 000 records, ~ 5.5Gb of
> data. pg_restore has been running for a week without any considerable
> progress. There are plenty of lines like these in the log:
> pg_restore: processing item 3125397 BLOB 10001967
> pg_restore: executing BLOB 10001967
> CPU usage is 100% always. The total database size is about 100 Gb and
> it restores in an hour or so without BLOBs.
In response to
pgsql-performance by date
|Next:||From: Tom Lane||Date: 2010-12-08 14:34:23|
|Subject: Re: Performance under contention |
|Previous:||From: Robert Haas||Date: 2010-12-08 04:24:14|
|Subject: Re: Performance under contention|
pgsql-hackers by date
|Next:||From: Dmitriy Igrishin||Date: 2010-12-08 08:51:05|
|Subject: Re: Feature request - CREATE TYPE ... WITH OID = oid_number.|
|Previous:||From: David Fetter||Date: 2010-12-08 08:19:49|
|Subject: Re: wCTE behaviour|