[Spacewalk-list] Sequences going wrong

Thomas Schweikle tschweikle at gmail.com
Tue Feb 19 09:30:09 UTC 2019


On 15.02.2019 18:19, Robert Paschedag wrote:
> Am 15. Februar 2019 09:28:26 MEZ schrieb Thomas Schweikle <tschweikle-Re5JQEeQqe8AvxtiuMwx3w at public.gmane.org>:
>> On 14.02.2019 12:09, P.Cookson-mHdZ94l+r7K1Qrn1Bg8BZw-XMD5yJDbdMReXY1tMh2IBg at public.gmane.org
>> wrote:
>>> Hi Thomas
>>>
>>> The "Internal Server Error" appears to be a really generic and
>> un-helpful message that comes out of Spacewalk in all sorts of
>> scenario's! I haven't experienced your particular problem but, in the
>> past, I've been advised to look at " /var/log/tomcat/catalina.out" to
>> try and get a better idea of what the issue is. Don't know if this will
>> help on this occasion?
>>
>> No, this was not helpful in this occasion. I'd need some way to make
>> spacewalk deliver what error was created, by taking logs from
>> /var/log/tomcat/catalina.out,
>> /var/lib/pgsql/11/data/log/postgres-*.log.
>>
>> It would be really helpful spacewalk pushing the error logs to the web
>> interface for debugging such cases!
>>
>>>
>>> Regards
>>> Phil
>>>
>>> -----Original Message-----
>>> From: spacewalk-list-bounces-H+wXaHxf7aLQT0dZR+AlfA-XMD5yJDbdMReXY1tMh2IBg at public.gmane.org
>> <spacewalk-list-bounces-H+wXaHxf7aLQT0dZR+AlfA-XMD5yJDbdMReXY1tMh2IBg at public.gmane.org> On
>> Behalf Of tschweikle-Re5JQEeQqe8AvxtiuMwx3w-XMD5yJDbdMReXY1tMh2IBg at public.gmane.org
>>> Sent: 14 February 2019 10:31
>>> To: spacewalk
>> <spacewalk-list-H+wXaHxf7aLQT0dZR+AlfA-XMD5yJDbdMReXY1tMh2IBg at public.gmane.org>
>>> Subject: [Spacewalk-list] Sequences going wrong
>>>
>>> Hi!
>>>
>>> We're having a lot of troubble with spacewalk loosing sequences, or
>>> better: loosing sync within sequences. As spacewalk runs all looks ok
>> for some time. then you may get "Internal server error" at a variety of
>> places. Mostly it starts showing up in "Schedule -> Actions" click on
>> one of the numbers behind in columns Finished, Failed, Pending.
>>> Instead of showing me details on where the action hav finished,
>> failed or is pending I am prompted with "Internal server error".
>>>
>>> Looking into the sql-logs I can find sequences out of sync. Fixing
>> these manually makes it work again -- for some time. I do not have any
>> idea why these sequences starting to get out of sync. Any idea? Bug?
>>>
>>> Spacewalk 2.9
>>> OS: CentOS 7
>>>
>>> --
>>> Thomas
>>>
>>> _______________________________________________
>>> Spacewalk-list mailing list
>>> Spacewalk-list-H+wXaHxf7aLQT0dZR+AlfA-XMD5yJDbdMReXY1tMh2IBg at public.gmane.org
>>> https://www.redhat.com/mailman/listinfo/spacewalk-list
>>>
> 
> Well... You can look into the stack traces at which file (and line) the error occurred. Clone the spacewalk git repo, checkout your version "branch" and look at the code.

That is what I am trying at the moment. BUT and its a really BIG BUT:
there some stack traces. And I need to have it related to what is going
on after I click some link. The way Spacewalk does it now I cant. I cant
relate a click to an error. Impossible. Absolutely impossible. Just
because the error does not show up with a fresh installation and three
clients. It shows up with 200 clients and a running, productive system.

I need Spacewalk to handle the traces thru to WebUI. How can I advice
tomcat to do this: write an error message, then write the full trace
right beyond.

This is the third time the database being corrupted this way. I DO NOT
want to start all over, taking about 400 clients, 800 configs copied by
cut and past to a new spacewalk to have the database corrupted 4th time
then WITHOUT ever finding what was the core cause!

-- 
Thomas




More information about the Spacewalk-list mailing list