Timeout - write_dma

Hi,

I recently installed FreeBSD 8.2 as a guest on my Xen Server 5.5.

I'm experiencing a lot of freezes for about 3 or 4 seconds and I get DMA Timeouts like these:
Code:
Jun 16 08:56:18 www1 kernel: ad0: TIMEOUT - WRITE_DMA retrying (1 retry left) LBA=7000596
Jun 16 08:57:19 www1 kernel: ad0: TIMEOUT - WRITE_DMA retrying (1 retry left) LBA=12314484
Jun 16 09:00:56 www1 kernel: ad0: TIMEOUT - READ_DMA retrying (1 retry left) LBA=12160756
Jun 16 09:01:21 www1 kernel: ad0: TIMEOUT - READ_DMA retrying (1 retry left) LBA=12160612
Jun 16 09:07:32 www1 kernel: ad0: TIMEOUT - WRITE_DMA retrying (1 retry left) LBA=4154996
Jun 16 09:13:31 www1 kernel: ad0: TIMEOUT - READ_DMA retrying (1 retry left) LBA=9086004

In that Xen Server I already have some other Linux machines (which use the same hard disk) and do not have that type of errors.

There's anything I can do? How can I check if the disk is really failing?

Thanks.
 
Old ata(4) subsystem (default on 8.x) has quite aggressive command timeouts. Can't it be that your physical disk is too busy to handle requests in a reasonable time? If your disk is indeed failing, you should check for it with the host OS, not with the guest.
 
The message I receive is:
Code:
timeout waiting for write DRQ
Is it related? I have an old model SATA controller Promise TX4 SATA300. I'm quite sure it is a driver issue, as I'm having the problem with FreeBSD 8.2-RELEASE-p1 but not with lower ones... Any advice?
 
I think not. I've seen number of reports from Promise TX4 SATA300 users. But without any docs and hardware I can't recommend something.
 
dum said:
In that thread I had written that I didn't have any major issues with that controller. Well, I was wrong, very wrong! Have a look at this relevant PR. To make a long story short, I have replaced three disks on that controller. The first one did indicate some issues with smartmon tools. The other two showed nothing. So, I suspected that there is something wrong with the controller.
 
Back
Top