[ovirt-users] Unexpected SSL errors (unexpected eof) in vdsm log?

[Piotr Kliczewski]

On Thu, Nov 10, 2016 at 5:18 PM, Derek Atkins <derek at ihtfp.com> wrote:
> Hi,
>
> On Thu, November 10, 2016 11:04 am, Piotr Kliczewski wrote:
> [snip]
>>> Could you explain the above error?  I'm not sure I understand what it
>>> means.  I do, however, know what the following error is about.  I
>>> attempted to import an OVA file and the file was mode 600 root:root and
>>> therefore was not readable by VDSM.  Hence the "Errno 13: Permission
>>> Denied":
>>>
>>
>> I added the stack traces to the email because I want someone from
>> storage and virt to take a look
>> at those failures. Maybe both were fixed already.
>
> The second stack trace, I presume, is due to my permission problem.  If so
>  then that's my own doing and not something to fix in oVirt.  (Unless my
> assumption is incorrect, but at the time the .ova file *was* unreadable).
>
>>>> org.ovirt.engine.core.vdsbroker.vdsbroker.VDSErrorException:
>>>> VDSGenericException: VDSErrorException: Failed to GetOvaInfoVDS, error
>>>> = [Errno 13] Permission denied:
>>>> u'/ovirt/import/openafs-fc23-64.ihtfp.org.ova', code = -32603 (Failed
>>>> with error unexpected and code 16)
>>>> at
>>> [snip]
>>>>
>>>> both not related to reconnects.
>>>
>>> Agreed.
>>>
>>>> I see that from time to time there are connections reset by peer
>>>>
>>>> 2016-11-04 10:58:43,442 ERROR
>>>> [org.ovirt.engine.core.vdsbroker.vdsbroker.GetStatsVDSCommand]
>>>> (DefaultQuartzScheduler6) [77387d45] Command
>>>> 'GetStatsVDSCommand(HostName = ovirt-0,
>>>> VdsIdAndVdsVDSCommandParametersBase:{runAsync='true',
>>>> hostId='62b75bb9-fbd9-405f-b479-b6ad8cffd5b1',
>>>> vds='Host[ovirt-0,62b75bb9-fbd9-405f-b479-b6ad8cffd5b1]'})' execution
>>>> failed: VDSGenericException: VDSNetworkException: Connection reset by
>>>> peer
>>>>
>>>> which means that vdsm or the host was stopped. Vdsm log you provided
>>>> do not cover this time so I am not able to say what is the cause of
>>>> it.
>>>
>>> Yeah, Nov 4 was approximately the time I was installing the systems, so
>>> yes, it's not surprising to see some up and down times around then.
>>>
>>>>
>>>> There are more 'unexpected eof' in the logs but they seems not to be
>>>> triggered by the engine. It looks like those connection are triggered
>>>> from local host.
>>>> This seems to be related to https://bugzilla.redhat.com/1349829
>>>
>>> Is there something I can do to test/verify this?
>>
>> It needs to be fixed first. I added it as reference for you.
>
> Thanks.  I was more asking if there is some way to verify that this *is*
> the bug I'm hitting?  Otherwise I have to wait for the fix for this bug
> and only then can I see if it actually fixes this issue or if it's
> something else.

I think we need to wait

>
> I'm not sure why not re-using the session would cause an EOF Error.
>
> Thanks,
>
> -derek
> --
>        Derek Atkins                 617-623-3745
>        derek at ihtfp.com             www.ihtfp.com
>        Computer and Internet Security Consultant
>