Hello Mark et al,
This is a mea culpa—it turns out we’re running firewalld on our “new” centos 7
systems and it was blocking port 389. If I disable firewalld I can’t repeat
the problem.
Thanks for your help on this.
-morgan
> On Aug 23, 2017, at 4:24 PM, Mark Reynolds <marey...@redhat.com> wrote:
>
>
>
> On 08/23/2017 03:09 PM, Morgan Jones wrote:
>> Mark,
>>
>> See attached. The break in the log is where it hung and then came back to
>> life.
> Odd, I don't see the connection being closed. Looks like the client
> (the console) did not send any requests during that time, what about the
> other servers at that same time?
>
> 23/Aug/2017:12:56:13 <--> 23/Aug/2017:12:59:58
>
> Also, did you try registering the servers in a different order?
>
> Perhaps try and get a stack trace from the console during the hang using
> jstack (from Oracle java)
>
> Thanks,
> Mark
>>
>> I’m also going to follow up with our networking and security folks to see if
>> we can find anything there. These hosts are all on the same subnet for what
>> it’s worth.
>>
>> Thanks for the help.
>>
>> -morgan
>>
>>
>>> On Aug 23, 2017, at 12:35 PM, Mark Reynolds <marey...@redhat.com> wrote:
>>>
>>>
>>>
>>> On 08/23/2017 12:31 PM, Morgan Jones wrote:
>>>>> On Aug 23, 2017, at 12:17 PM, Mark Reynolds <marey...@redhat.com> wrote:
>>>>>
>>>>>
>>>>>> [pid 27442] recvmsg(14, 0x7f3880ef74d0, 0) = -1 EAGAIN (Resource
>>>>>> temporarily unavailable)
>>>>>> [pid 27442] recvmsg(14, 0x7f3880ef74d0, 0) = -1 EAGAIN (Resource
>>>>>> temporarily unavailable)
>>>>>> [pid 27442] poll([{fd=14, events=POLLRDNORM}, {fd=15,
>>>>>> events=POLLRDNORM}], 2, 500 <unfinished ...>
>>>>>> [pid 27440] <... futex resumed> ) = -1 ETIMEDOUT (Connection timed
>>>>>> out)
>>>>>> [pid 27440] futex(0x7f38940cfd28, FUTEX_WAKE_PRIVATE, 1) = 0
>>>>> Sorry forgot to comment on this...
>>>>>
>>>>> This explains the "hang" - connections to the remove server(s) are
>>>>> timing out.
>>>>>
>>>>> Can you look at the DS access logs on a remote server during the hang
>>>>> (note there is a 30 sec log buffer with the access log). Perhaps just
>>>>> tail the access log, reproduce the hang (wait 30 seconds), and provide
>>>>> the complete tail output.
>>>> Aha, ok, thanks. Which access log do you want or do you want all of them?
>>> Just the access log from the server that triggers the initial hang.
>>>> -morgan
>
_______________________________________________
389-users mailing list -- 389-users@lists.fedoraproject.org
To unsubscribe send an email to 389-users-le...@lists.fedoraproject.org
