Better way to detect failed disks

[zerebubuth]

Recently we've lost disks in ramoth and yevaud without notifications (at least - not that I can find). However, failed disk notifications appear to be working fine on orm and pummelzacken.

[tomhughes]

It's basically a function of which RAID controller is in use - if it's soft raid or a hardware RAID that http://hwraid.le-vert.net/ has support for then we get notifications. If it's a hardware RAID with only proprietary tools and no wrapper script then we don't.

[tomhughes]

Looks like ramoth should be able to detect them, but the status command is bombing out:

ramoth [~] % sudo megaclisas-status 
[sudo] password for tomh: 
-- Controller informations --
-- ID | Model
c0 | LSI MegaRAID SAS 9271-8iCC

-- Arrays informations --
-- ID | Type | Size | Status | InProgress
c0u0 | RAID1 | 465G | Optimal | None
c0u1 | RAID6 | 7633G | Optimal | None
Traceback (most recent call last):
  File "/usr/sbin/megaclisas-status", line 164, in <module>
    arrayinfo = returnArrayInfo(output,controllerid,arrayid)
  File "/usr/sbin/megaclisas-status", line 101, in returnArrayInfo
    if ldpdcount and (int(spandepth) > 1):
UnboundLocalError: local variable 'ldpdcount' referenced before assignment

[tomhughes]

The yevaud card is an Areca which hwraid doesn't have a status daemon for.

[tomhughes]

The error on ramoth seems to be because megacli is claiming there are four arrays but will only return information about the first one.

[tomhughes]

Ah I was wrong about that, the actual problem is array 2, which doesn't report a number of disks:

ramoth [~] % sudo megacli -LDInfo -l2 -a0 -NoLog


Adapter 0 -- Virtual Drive Information:
CacheCade Virtual Drive: 2 (Target Id: 2)
Virtual Drive Type    : CacheCade 
Name          : 
RAID Level        : Primary-1, Secondary-0
State             : Optimal
Size          : 558.406 GB
Target Id of the Associated LDs : 0,3,1
Default Cache Policy  : WriteBack, ReadAdaptive, Direct, No Write Cache if Bad BBU
Current Cache Policy  : WriteBack, ReadAdaptive, Direct, No Write Cache if Bad BBU



Exit Code: 0x00

[tomhughes]

I've patched it now but it seems to be reporting everything as optimal anyway.

[zerebubuth]

Indeed - the arrays are optimal. But there's a global hot-spare, which has been pulled into the array to stand in for the failed drive. So there's a failed physical disk, which isn't part of any array... but we still want to know about it.

[tomhughes]

So send a patch to https://github.com/eLvErDe/hwraid then ;-)

[tomhughes]

eLvErDe/hwraid#13 may actually be what we want?

[tomhughes]

I've opened eLvErDe/hwraid#17 for the CacheCade issue.

[zerebubuth]

Is this still an issue, or are we getting much better alerting these days?

[Firefishy]

Closing as very old. If still an issue can be re-opened.