Mail server frequent down

[redshadow]

Can anyone tell what does
hda: lost interrupt mean? and how can we fix this.

Our mail server is in this server which always gives a message hda: lost interrupt.

What we do is always to restart sendmail or restart linux.
When rebooted it will proceed to fsck and finds errors which it automatically corrected.

Thanks in advance, we're up to this for several days now.

 

[marsd]

Lost interrupt is a bad sign:
means that a disk op is timing out or not being completed.
I had that with an old cdrom drive before it went bad,
and also encountered it one time with a misconfigured
pio only ide drive configured for dma with hdparm.

Sounds like you need to backup soon and look at
where you need to start your tshooting of the issue.
I'd just replace the buggy drive, and RIP the old one.

 

[redshadow]

maybe I should take your advice immediately.

I just can't figure out if this means there is wrong configuration with the software or is it purely a hardware failure itself.

Thanks marsd.

 

[ifincham]

Hi,

As hinted at by marsd, you could check the dma (direct memory access) settings of the drive :

/sbin/hdparm -v /dev/hda

If dma is on you can try turning it off as follows :

/sbin/hdparm -d0 /dev/hda

To preserve the latter across reboots you'd need to put a line like that in /etc/rc.d/rc.sysinit .

See also -->

http://www.pcquest.com/archive/articledetail.asp?arid=11733

Regards