dhis2-users team mailing list archive
-
dhis2-users team
-
Mailing list archive
-
Message #13682
Major speed issue with pg_restore
-
To:
"dhis2-users@xxxxxxxxxxxxxxxxxxx" <dhis2-users@xxxxxxxxxxxxxxxxxxx>
-
From:
Edward Robinson <erobinson@xxxxxxxxxxxxxxxxxx>
-
Date:
Fri, 14 Jul 2017 01:54:52 +0000
-
Accept-language:
en-US
-
Authentication-results:
lists.launchpad.net; dkim=none (message not signed) header.d=none;lists.launchpad.net; dmarc=none action=none header.from=projectbalance.com;
-
Spamdiagnosticmetadata:
NSPM
-
Spamdiagnosticoutput:
1:99
-
Thread-index:
AdL8Qv0hVdlHNLGQToSpRSLZ4Y1GzQ==
-
Thread-topic:
Major speed issue with pg_restore
I've setup a new Ubuntu 16.04 box with PostgreSQL 9.5.5 and I'm restoring a fairly large DHIS2 backup but having speed issues. It's a full pg_dump in custom format and about 650Gb compressed (a plain text dump produces a 7Gb file). I made sure inserts were turned off, so that's not the issue, but so far it's been running for 33 hours - CPU at 100% - with no end in sight. This is a backup file that took 30 minutes to generate.
If it is running synchronously, I calculated that it's on around 20 million DB rows of around 170 million after 33 hours! Surely that's not normal.
I've tweaked PostgreSQL with the following settings:
maintenance_work_mem=2GB
max_wal_size = 1Gb
checkpoint_timeout = 3600
checkpoint_completion_target = 0.9
Anyway, if anyone has insight or has had a similar experience, or suggestions, please let me know!
I'm testing it on another (Windows) instance to see if there's something amiss.
Cheers,
Ed
Follow ups