Lengthy pg_restore and high iowait? - Mailing list pgsql-performance
From | Leon Out |
---|---|
Subject | Lengthy pg_restore and high iowait? |
Date | |
Msg-id | 103862F9-5E68-11D8-AB22-0030658FB514@comvision.com Whole thread Raw |
List | pgsql-performance |
<fixed><fontfamily><DEFANGED_param>Courier New</DEFANGED_param>Hello again. I'm setting up a backup routine for my new db server. As part of my testing, I'm attempting to pg_restore a pg_dump'ed backup of my database. The database is about 4.3 GB, and the dump file is about 100 MB. I first did a schema-only restore, then started a data-only restore with --disable-triggers to get around the referential integrity issues of reloading the data. The data-only restore has been running for a couple of hours now, and I'm seeing high iowait numbers in top. 15:57:58 up 23:55, 2 users, load average: 2.04, 2.07, 2.01 60 processes: 57 sleeping, 3 running, 0 zombie, 0 stopped CPU states: cpu user nice system irq softirq iowait idle total 4.0% 0.0% 0.7% 0.0% 0.0% 43.5% 51.6% cpu00 0.0% 0.0% 0.3% 0.0% 0.0% 84.8% 14.7% cpu01 15.7% 0.0% 1.7% 0.0% 0.0% 2.7% 79.6% cpu02 0.1% 0.0% 0.7% 0.0% 0.0% 84.2% 14.7% cpu03 0.2% 0.0% 0.0% 0.0% 0.0% 2.4% 97.4% Mem: 3869544k av, 3849280k used, 20264k free, 0k shrd, 110544k buff 1297452k actv, 2298928k in_d, 57732k in_c Swap: 2040244k av, 0k used, 2040244k free 3576684k cached PID USER PRI NI SIZE RSS SHARE STAT %CPU %MEM TIME CPU COMMAND 8366 postgres 16 0 954M 954M 952M R 4.3 25.2 21:58 1 postmaster 9282 postgres 15 0 0 0 0 RW 0.2 0.0 0:00 2 postmaster 1 root 15 0 496 496 444 S 0.0 0.0 0:05 3 init Questions: * Do these iowait numbers indicate a problem with my setup? * Does anyone have a good method for disabling indexes before a restore and restoring them afterwards? I've spent some time writing scripts to do this, but I have yet to come up with drop/recreate solution that returns my database to the same pre-drop state. Thanks in advance! Leon</fontfamily></fixed> Hello again. I'm setting up a backup routine for my new db server. As part of my testing, I'm attempting to pg_restore a pg_dump'ed backup of my database. The database is about 4.3 GB, and the dump file is about 100 MB. I first did a schema-only restore, then started a data-only restore with --disable-triggers to get around the referential integrity issues of reloading the data. The data-only restore has been running for a couple of hours now, and I'm seeing high iowait numbers in top. 15:57:58 up 23:55, 2 users, load average: 2.04, 2.07, 2.01 60 processes: 57 sleeping, 3 running, 0 zombie, 0 stopped CPU states: cpu user nice system irq softirq iowait idle total 4.0% 0.0% 0.7% 0.0% 0.0% 43.5% 51.6% cpu00 0.0% 0.0% 0.3% 0.0% 0.0% 84.8% 14.7% cpu01 15.7% 0.0% 1.7% 0.0% 0.0% 2.7% 79.6% cpu02 0.1% 0.0% 0.7% 0.0% 0.0% 84.2% 14.7% cpu03 0.2% 0.0% 0.0% 0.0% 0.0% 2.4% 97.4% Mem: 3869544k av, 3849280k used, 20264k free, 0k shrd, 110544k buff 1297452k actv, 2298928k in_d, 57732k in_c Swap: 2040244k av, 0k used, 2040244k free 3576684k cached PID USER PRI NI SIZE RSS SHARE STAT %CPU %MEM TIME CPU COMMAND 8366 postgres 16 0 954M 954M 952M R 4.3 25.2 21:58 1 postmaster 9282 postgres 15 0 0 0 0 RW 0.2 0.0 0:00 2 postmaster 1 root 15 0 496 496 444 S 0.0 0.0 0:05 3 init Questions: * Do these iowait numbers indicate a problem with my setup? * Does anyone have a good method for disabling indexes before a restore and restoring them afterwards? I've spent some time writing scripts to do this, but I have yet to come up with drop/recreate solution that returns my database to the same pre-drop state. Thanks in advance! Leon
pgsql-performance by date: