Question: Have the network configuration of OpenCSW been changed?

[Dagobert Michelsen]

Hi Yusuke,

> Am 03.12.2019 um 10:38 schrieb Yusuke Endoh <mame at ruby-lang.org>:
> 
> Thank you for the answer!
> 
> I suspected the network configuration because:
> 
> * the failures look related to network issue,
> * all OpenCSW machines started failing all at once,
> * the failures have never occurred on any other machines, and
> * I couldn't find any relevant commits whose timestamp is near to that timing.

There have been similar errors in the past which disappeared when I added IPv6 localhost,
but this was added from the start after the reboot, so this is probably not the issue.

> I don't know what the tests do :-)
> To be honest, debugging in OpenCSW is painful because it is slow, so I wanted to isolate the problem before debugging.

I suggest to start with the *x machine which should be reasonably fast as x86 on our VMware farm,
then probably 11s which is a moderately new M3000. The *10s are the slowest ones.


Best regards

  — Dago

> But I understand there has been no network change, so I'll give it a try to debug.  Thank you very much!
> 
> Best regards,
> 
> 2019年12月3日(火) 17:15 Dagobert Michelsen <dam at opencsw.org <mailto:dam at opencsw.org>>:
> Hi Yusuke,
> 
> Am 03.12.2019 um 00:41 schrieb Yusuke Endoh <mame at ruby-lang.org <mailto:mame at ruby-lang.org>>:
>> I'm Yusuke Endoh, a member of Ruby development team.
>> 
>> Since 2 A.M. 1st Dec. 2019 UTC, our Ruby CIs on OpenCSW have failed.
>> The failure looks related to network configuration about localhost.
>> 
>> https://rubyci.org/logs/rubyci.s3.amazonaws.com/unstable10s/ruby-master/log/20191202T171915Z.fail.html.gz <https://rubyci.org/logs/rubyci.s3.amazonaws.com/unstable10s/ruby-master/log/20191202T171915Z.fail.html.gz>
>> 
>> ```
>> DRb::DRbConnError: drbssl://::1:35909 <> - #<Errno::ECONNREFUSED: Connection refused - connect(2) for "::1" port 35909>
>> ```
>> 
>> So, I'd like to ask if something network stuff was changed?
> 
> We rebooted the machine and some of the networking setup doesn’t come up automatically due to
> routing issues between the global zone and local zones. However, I just looked and as far as I can
> say it looks good to me.
> 
>> If it is an intentional change, I'd be happy to tackle to fix the issue on the side of Ruby test suite.
>> I have seen the issue on unstable10s, 10x, 11s, and 11x.
> 
> unstable10x, unstable11s and unstable11x have not been rebooted, these are different machines,
> so if you see the error there too it is not related due to networking on the rebooted host.
> 
>> FWIW, `nslookup localhost` fails to find 127.0.0.1.
>> 
>> ```
>> rubyci at unstable10s [unstable10s]:~ > nslookup localhost                                                
>> Server:         192.168.1.6
>> Address:        192.168.1.6#53
>> 
>> ** server can't find localhost: NXDOMAIN
>> ```
> 
> It is our normal configuration to not resolve localhost via DNS, but you can resolve it locally
> with the resolver lib:
> 
> root at unstable10s [unstable10s]:/root > getent hosts localhost 
> 127.0.0.1       localhost
> root at unstable10s [unstable10s]:/root > nslookup localhost
> Server:         192.168.1.6
> Address:        192.168.1.6#53
> 
> ** server can't find localhost: NXDOMAIN
> 
> zsh: 26934 exit 1     nslookup localhost
> root at unstable10s [unstable10s]:/root > getent hosts localhost
> 127.0.0.1       localhost
> root at unstable10s [unstable10s]:/root > 
> 
> 
> Maybe you can describe what the test actually does and why it technically fails,
> I am not sure at the moment what has changed, be it the build machine configuration
> or something in your code or something else.
> 
> 
> Best regards
> 
>   — Dago
> 
> 
> -- 
> "You don't become great by trying to be great, you become great by wanting to do something,
> and then doing it so hard that you become great in the process." - xkcd #896
> 

-- 
"You don't become great by trying to be great, you become great by wanting to do something,
and then doing it so hard that you become great in the process." - xkcd #896

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.opencsw.org/pipermail/buildfarm/attachments/20191203/621579a2/attachment-0001.html>