| From: | Scott Marlowe <scott(dot)marlowe(at)gmail(dot)com> |
|---|---|
| To: | Richard Yen <richyen(at)iparadigms(dot)com> |
| Cc: | pgsql-performance(at)postgresql(dot)org, slony List <slony1-general(at)lists(dot)slony(dot)info> |
| Subject: | Re: WAL partition overloaded--by autovacuum? |
| Date: | 2010-07-07 03:25:28 |
| Message-ID: | AANLkTik1oR2fMBp9pzpeYUgrANXKpK0B23scXONQEK1z@mail.gmail.com |
| Views: | Raw Message | Whole Thread | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-performance |
On Tue, Jul 6, 2010 at 7:10 PM, Richard Yen <richyen(at)iparadigms(dot)com> wrote:
> This leads me to believe that there was a sudden flurry of write activity that occurred, and the process that would flush WAL files to /db/data/ couldn't keep up, thereby filling up the disk. I'm wondering if anyone else out there might be able to give me some insight or comments to my assessment--is it accurate? Any input would be helpful, and I'll try to make necessary architectural changes to keep this from happening again.
I tend to agree. What kind of disk setup is under your data store?
If it's RAID-5 or RAID-6 is it possibly degraded?
I'd run iostat -xd 10 while the write load it high to see how your
xlog partition %util compares to the main data store partition.
Do you have your main data store partition under a battery backed
caching RAID controller? Tell us what you can about your hardware
setup.
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Scott Marlowe | 2010-07-07 03:27:10 | Re: WAL partition overloaded--by autovacuum? |
| Previous Message | Richard Yen | 2010-07-07 01:10:10 | WAL partition overloaded--by autovacuum? |