[OmniOS-discuss] Multiple faulty SSD's ?

[Shaun McGuane]

Thanks everyone for you advice on this.

What a pain in the ass…..

Shaun


From: OmniOS-discuss [mailto:omnios-discuss-bounces at lists.omniti.com] On Behalf Of Schweiss, Chip
Sent: Wednesday, 27 July 2016 5:55 AM
To: Piotr Jasiukajtis <estibi at me.com>
Cc: omnios-discuss <omnios-discuss at lists.omniti.com>
Subject: Re: [OmniOS-discuss] Multiple faulty SSD's ?

I don't have a lot of experience with the 850 Pro, but a lot with the 840 Pro under OmniOS

With 4K block size set in sd.conf and slicing them to only use 80% of their capacity a pool of 72 of them has been under near constant heavy read/write workload for over 3 years without a single chksum error.

-Chip

On Tue, Jul 26, 2016 at 1:30 PM, Piotr Jasiukajtis <estibi at me.com<mailto:estibi at me.com>> wrote:
I don’t know a root cause, but it’s better to have a workaround than a corrupted pools.

--
Piotr Jasiukajtis

> On 26 Jul 2016, at 20:06, Dan McDonald <danmcd at omniti.com<mailto:danmcd at omniti.com>> wrote:
>
> I wonder if those sd.conf changes should be upstreamed or not?
>
> Dan
>
> Sent from my iPhone (typos, autocorrect, and all)
>
>> On Jul 26, 2016, at 1:28 PM, Piotr Jasiukajtis <estibi at me.com<mailto:estibi at me.com>> wrote:
>>
>> You may want to force the driver to use 4k instead of 512b for those drivers and create a new pool:
>>
>> https://github.com/joyent/smartos-live/commit/dd25937d2f9725def16f5e8dbb16a8bcbc2213d5
>>
>> --
>> Piotr Jasiukajtis
>>
>>> On 26 Jul 2016, at 02:24, Shaun McGuane <shaun at rackcentral.com<mailto:shaun at rackcentral.com>> wrote:
>>>
>>> Hi List,
>>>
>>> I want to report very strange SSD behaviour on a new pool I setup.
>>>
>>> The hardware is a HP DL180 G6 Server with the LSI 9207-8i Card
>>> And 8x 1TB Samsung SSD Pro drives. Running omnios-10b9c79
>>>
>>> All the drives are brand spanking new setup in a raidz2 array.
>>>
>>> Within 2 months the below has happened and there has been very
>>> Little use on this array.
>>>
>>> pool: SSD-TANK
>>> state: DEGRADED
>>> status: One or more devices are faulted in response to persistent errors.
>>>       Sufficient replicas exist for the pool to continue functioning in a
>>>       degraded state.
>>> action: Replace the faulted device, or use 'zpool clear' to mark the device
>>>       repaired.
>>> scan: scrub repaired 23K in 1h12m with 0 errors on Mon Jul 25 20:13:04 2016
>>> config:
>>>
>>>       NAME                       STATE     READ WRITE CKSUM
>>>       SSD-TANK                   DEGRADED     1    67    35
>>>         raidz2-0                 DEGRADED     4    72   113
>>>           c5t500253884014D0D3d0  ONLINE       0     0     2
>>>           c5t50025388401F767Ad0  DEGRADED     0     0    19  too many errors
>>>           c5t50025388401F767Bd0  FAULTED      0     0     0  too many errors
>>>           c5t50025388401F767Dd0  ONLINE       0     0     0
>>>           c5t50025388401F767Fd0  ONLINE       0     0     1
>>>           c5t50025388401F7679d0  ONLINE       0     0     2
>>>           c5t50025388401F7680d0  REMOVED      0     0     0
>>>           c5t50025388401F7682d0  ONLINE       0     0     1
>>>
>>> Can anyone suggest why I would have this problem where I am seeing CKSUM errors
>>> On most disks and while only one has faulted others have been degraded or removed.
>>>
>>> Thanks
>>> Shaun
>>> _______________________________________________
>>> OmniOS-discuss mailing list
>>> OmniOS-discuss at lists.omniti.com<mailto:OmniOS-discuss at lists.omniti.com>
>>> http://lists.omniti.com/mailman/listinfo/omnios-discuss
>>
>> _______________________________________________
>> OmniOS-discuss mailing list
>> OmniOS-discuss at lists.omniti.com<mailto:OmniOS-discuss at lists.omniti.com>
>> http://lists.omniti.com/mailman/listinfo/omnios-discuss

_______________________________________________
OmniOS-discuss mailing list
OmniOS-discuss at lists.omniti.com<mailto:OmniOS-discuss at lists.omniti.com>
http://lists.omniti.com/mailman/listinfo/omnios-discuss

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://omniosce.org/ml-archive/attachments/20160727/b47ead78/attachment.html>