[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: megaraid and 2.6 kernel



Maarten Vink said:
> R. W. Rodolico wrote:
>> There are some very interesting things in /proc/scsi, but I don't
>> know what all of them are. If you tell me the name of some tools,
>> I'll be happy to try them. This box doesn't go into production for
>> a
>> few more days, and I can rebuild rapidly if I mess something up.
>>
>> I went into /proc/scsi/sg/devices and played around. Did a
>> cat device_hdr devices device_strs
>> and got
>>
>> host    chan    id      lun     type    opens   qdepth  busy
>> online
>> 0       0       6       0       3       1       64      0       1
>> 0       2       0       0       0       1       64      0       1
>> PE/PV           1x6 SCSI BP             1.0
>> MegaRAID        LD 0 RAID1  139G        514K
>>
>> but this all appears to be talking just about the card, not the
>> attached drives. I pulled a drive, then plugged it back in, just
>> for
>> fun, and the above values did not change (I was hoping busy or
>> online would). I know there is something, somewhere, 'cause during
>> boot a lot of interesting info about drives is placed into syslog.
>> Just don't see any of it in /proc
>>
>> Rod
>
> The tools most likely to work are those at
> http://www.lsilogic.com/downloads/downloads.do?product=8069
>
> All the way towards the bottom of the page there are a "MegaMGR" and
> "GAM" for
> linux. We have had the GAM working om a 2.4 kernel with the standard
> megaraid
> driver, but we didn't have any luck with 2.6.
>
> How were you planning to monitor the status of your RAID arrays?
> Having a RAID
> array is no use if more than one drive fails, and we don't regularly
> check our
> server's status displays for RAID errors.
>
> Maarten
>
>
> --
> To UNSUBSCRIBE, email to debian-isp-REQUEST@lists.debian.org
> with a subject of "unsubscribe". Trouble? Contact
> listmaster@lists.debian.org
>
>

I'll jump out there and see about the monitoring tool. Thanks.

Monitoring the RAID is easy. The server room is right next to an
area that has a lot of workers 18 hours a day. The controller emits
a very intense beep when errors occur, loud enough to cut through
the sound of the other half dozen machines running. I figure the
workers will be coming to me to get me to turn off the loud,
irritating noise when/if the RAID fails. Basic biological monitoring
system :)

Actually, I want the monitoring tools so I can be at the server
BEFORE anyone gets horribly irritated.

Rod


-- 
Meddle not in the Affairs of Dragons
    for thou art crunchy, and good with catsup.




Reply to: