Disk Enclosures
1752788 Members
6105 Online
108789 Solutions
New Discussion юеВ

Re: K380 w 12H dies after diskfailure

 
Donald Kok
Respected Contributor

K380 w 12H dies after diskfailure

Hi,

The other week a disk in the 12H failed. There were continuous scsi errors in the syslog. The frontpanel showed 12H, and the filesystems on the 12H were not accessible. This all solved when the new disk was balanced in the 12H.

This behaviour is not what I expected from an AutoRaid. I expected the system would work as usual when a disk fails.

Thanks in Advance
Donald
My systems are 100% Murphy Compliant. Guaranteed!!!
9 REPLIES 9
Ricardo Rocha
Valued Contributor

Re: K380 w 12H dies after diskfailure

Hi

Sometimes when a disk fails, it fills the bus with "garbage", ie instead of being dead, it sends lots of noise in the bus. This way, the host cannot reach the rest of the disks of the same bus. That's why your host probably couldn't see the filesystem in the 12h.

Bye
"there is this old man who spent so much of his life sleeping that he is able to keep awake for the rest of his years"
Bill McNAMARA_1
Honored Contributor

Re: K380 w 12H dies after diskfailure

the internal busses on the autoraid, that the disks are on are seperate from the hosts fw bus.
you should not see this.

arraydsp -a

logprint (see man)

and
strings /etc/lvmtab
mount -p
vgdisplay vg_s_on_12H
ioscan -fnk

attach in response.

Later,
Bill
It works for me (tm)
Donald Kok
Respected Contributor

Re: K380 w 12H dies after diskfailure

Hi Bill,

Here they are....

My systems are 100% Murphy Compliant. Guaranteed!!!
Donald Kok
Respected Contributor

Re: K380 w 12H dies after diskfailure

Hey, that's funny, man can only attach 1 file. Let's tar.
Greetzz
Donald
My systems are 100% Murphy Compliant. Guaranteed!!!
Donald Kok
Respected Contributor

Re: K380 w 12H dies after diskfailure

and again....

My systems are 100% Murphy Compliant. Guaranteed!!!
Donald Kok
Respected Contributor

Re: K380 w 12H dies after diskfailure

upload of tar and gz file can not be send (communication lost by server), ordinary files can be send.

That were vgdisplay and mount, this is lvmtab.
My systems are 100% Murphy Compliant. Guaranteed!!!
Donald Kok
Respected Contributor

Re: K380 w 12H dies after diskfailure

logprint not allowed too
this is ioscan
My systems are 100% Murphy Compliant. Guaranteed!!!
Donald Kok
Respected Contributor

Re: K380 w 12H dies after diskfailure

Hi bill,
arraydsp also not allowed.

Should I send you an email, or do you know a better way to attach these files?

Greetzz
Donald
My systems are 100% Murphy Compliant. Guaranteed!!!
Steve Labar
Valued Contributor

Re: K380 w 12H dies after diskfailure

I have seen SCSI Timeouts and errors in my syslog before due to a bad disk in a 12H. When I first saw it, I thought I had a bad controller because of the SCSI ID reported, until I did some more checking with arraydsp and logprint. In my case, the disk did not have a catastrophic failure, but it did have enough bad sectors that it was trying to correct that it slowed the internal bus down, which in turn slowed the controller response down. Replacing the bad disk did clear everything up for me.

Hope this helps.

Steve