[OmniOS-discuss] [discuss] COMSTAR hanging

[Brian Hechinger]

ZFS lockups don’t seem much better an option. :)

Hardware is a Xeon box with 6x SATA disks in a raid10 wired straight to the controller. No expanders.

This is not a heavily loaded system (yet) so I don’t think it’s a load issue, sadly.

I *might* be able to downgrade to 014? I’m not sure I want to. I’d rather help get things fixed going forward.

This is for my home VMware stack. I have important services on disks local to the ESXi hosts (not idea, but makes this less painful when it happens) so COMSTAR locking up is mostly an inconvenience yet at this point. I’d rather it didn’t though because I’d like to be using this more heavily than it is now.

Thanks for the data points!

-brian

> On Jan 8, 2016, at 9:36 AM, Matej Zerovnik <matej at zunaj.si> wrote:
> 
> I had the same problems… 
> 
> In my case, comstart hanging went away with downgrade to early 014 version, but then ZFS started to lock.
> 
> What is your hardware? Any JBODs? SAS or SATA drives? Expanders?
> 
> Currently, I didn’t had Comstar lock for about a month, and I’m running latest 014 (but I did reduce the number of users for about 25%, so maybe I removed some some of the problematic users).
> 
> Matej
> 
> 
>> On 06 Jan 2016, at 21:29, Brian Hechinger <wonko at 4amlunch.net <mailto:wonko at 4amlunch.net>> wrote:
>> 
>> Great, look for one in the future when this happens again. :)
>> 
>> -brian
>> 
>>> On Jan 5, 2016, at 11:45 PM, Garrett D'Amore <garrett at damore.org <mailto:garrett at damore.org>> wrote:
>>> 
>>> Actually, what is probably the most useful is this command:
>>> 
>>> # echo ‘$<threadlist’ | mdb -k
>>> 
>>> A full crashdump will have that inside it, as well, but that first list of threads (and therefore will include the comstar threads) and backtraces will probably yield the most fruit for least effort on your part.
>>> 
>>> On Tue, Jan 5, 2016 at 6:59 PM, Brian Hechinger <wonko at 4amlunch.net <mailto:wonko at 4amlunch.net>> wrote:
>>> So this is the second time this has happened to me. The COMSTAR layer appears to be getting hung. At first I thought it was just the IB/SRP target stuff, but the iSCSI target also stops working. So far the only solution I’ve found is a reboot.
>>> 
>>> This is very concerning and I’d like to try to get it figured out.
>>> 
>>> The next time it happens, what is the best course of action in order to get the information you all need to debug this? I’m assuming force a crashdump, but is there anything else I could be doing?
>>> 
>>> Thanks!
>>> 
>>> -brian
>>> 
>>> PS: Latest OmniOS-stable
>>> 
>>> -------------------------------------------
>>> illumos-discuss
>>> Archives: https://www.listbox.com/member/archive/182180/=now <https://www.listbox.com/member/archive/182180/=now>
>>> RSS Feed: https://www.listbox.com/member/archive/rss/182180/22003744-9012f59c <https://www.listbox.com/member/archive/rss/182180/22003744-9012f59c>
>>> Modify Your Subscription: https://www.listbox.com/member/?member_id=22003744&id_secret=22003744-e9cd8436 <https://www.listbox.com/member/?member_id=22003744&id_secret=22003744-e9cd8436>
>>> Powered by Listbox: http://www.listbox.com <http://www.listbox.com/>
>>> 
>>> _______________________________________________
>>> 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 <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/20160108/cbb5d678/attachment-0001.html>