Date: Thu, 9 Mar 2006 11:03:40 +0000 From: Ashley Moran <work@ashleymoran.me.uk> To: freebsd-amd64@freebsd.org Subject: out of memory in ata_raid_init_request Message-ID: <200603091103.40224.work@ashleymoran.me.uk>
next in thread | raw e-mail | index | archive | help
I've got a server with an Asus A8V Deluxe board and two Seagate Barracuda 7200.9 disks plugged into the VIA controller in RAID 1. Every now and again I get errors like this: Mar 9 04:10:51 don FAILURE - out of memory in ata_raid_init_request Mar 9 04:10:51 don FAILURE - out of memory in ata_raid_init_request Mar 9 04:10:51 don FAILURE - out of memory in ata_raid_init_request Mar 9 04:10:51 don Fg_vfs_done():ar0s1h[WRITE(offset=87667261440, length=16384)]error = 5 Mar 9 04:10:51 don g_vfs_done():ar0s1h[WRITE(offset=87667343360, length=32768)]error = 5 Mar 9 04:10:51 don g_vfs_done():ar0s1h[WRITE(offset=87667474432, length=16384)]error = 5 Mar 9 04:10:51 don AILURE - out of memory in ata_raid_init_request Mar 9 04:10:51 don g_vfs_done():ar0s1h[WRITE(offset=87668228096, length=16384)]error = 5 I posted this to questions@ last time it happens but nobody could offer any advice. Is it hardware-specific, is it significant, and is it another reason to switch to gmirror? It is likely caused by postgres as that would have been doing a data import task around the time of the error. I haven't yet configured postgres to log anything useful, but the last entry in /var/log/messages was Mar 9 04:15:21 don postgres[70314]: [1-1] ERROR: tuple concurrently updated Ashley
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?200603091103.40224.work>