tracker.ceph.com

tracker.ceph.com

[Dan Mick]

tracker.ceph.com is having issues. I'm looking at it.

Re: tracker.ceph.com

[David Galloway]

On 12/17/2016 05:25 PM, Dan Mick wrote:
> tracker.ceph.com is having issues. I'm looking at it.
>
> --

The Tracker VM's disk was corrupted during an Openstack live migration.

A new VM was created and the Tracker data has been restored from a 
backup dated 12-17-2016 07:30 UTC.  Any changes made to bugs after that 
time are gone.

The Agile plugin has been restored and file attachments have been 
rsynced to the new VM.

Please let me know if you notice anything is amiss.

Re: tracker.ceph.com

[Nathan Cutler]

> Please let me know if you notice anything is amiss.

I haven't received any email notifications since the crash. Normally on 
a Monday I'd have several dozen.

-- 
Nathan Cutler
Software Engineer Distributed Storage
SUSE LINUX, s.r.o.
Tel.: +420 284 084 037

Re: tracker.ceph.com

[David Galloway]

On 12/19/2016 11:24 PM, Nathan Cutler wrote:
>> Please let me know if you notice anything is amiss.
>
> I haven't received any email notifications since the crash. Normally on
> a Monday I'd have several dozen.
>

Looks like it was trying to send mail over IPv6 and failing.

I switched back to postfix, disabled IPv6, and show a message was 
recently queued for delivery to you.  Please confirm you got it.

Re: tracker.ceph.com

[Nathan Cutler]

> Looks like it was trying to send mail over IPv6 and failing.
>
> I switched back to postfix, disabled IPv6, and show a message was
> recently queued for delivery to you.  Please confirm you got it.

Got it. Thanks for the fix!

-- 
Nathan Cutler
Software Engineer Distributed Storage
SUSE LINUX, s.r.o.
Tel.: +420 284 084 037

Re: tracker.ceph.com

[Wido den Hollander]

> Op 20 december 2016 om 17:31 schreef Nathan Cutler <ncutler-AlSwsSmVLrQ@public.gmane.org>:
> 
> 
> > Looks like it was trying to send mail over IPv6 and failing.
> >
> > I switched back to postfix, disabled IPv6, and show a message was
> > recently queued for delivery to you.  Please confirm you got it.
> 
> Got it. Thanks for the fix!
> 

I wouldn't call it a fix. In 2016 (almost 2017) IPv6 should be enabled, no questions asked.

After the recent migrations both ceph.com and tracker.ceph.com lost their IPv6. To me that's broken connectivity.

When is this coming back?

Wido

> -- 
> Nathan Cutler
> Software Engineer Distributed Storage
> SUSE LINUX, s.r.o.
> Tel.: +420 284 084 037
> _______________________________________________
> ceph-users mailing list
> ceph-users-idqoXFIVOFJgJs9I8MT0rw@public.gmane.org
> http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com

Re: [ceph-users] tracker.ceph.com

[David Galloway]

On 12/20/2016 11:48 AM, Wido den Hollander wrote:
>
>> Op 20 december 2016 om 17:31 schreef Nathan Cutler <ncutler@suse.cz>:
>>
>>
>>> Looks like it was trying to send mail over IPv6 and failing.
>>>
>>> I switched back to postfix, disabled IPv6, and show a message was
>>> recently queued for delivery to you.  Please confirm you got it.
>>
>> Got it. Thanks for the fix!
>>
>
> I wouldn't call it a fix. In 2016 (almost 2017) IPv6 should be enabled, no questions asked.
>
> After the recent migrations both ceph.com and tracker.ceph.com lost their IPv6. To me that's broken connectivity.
>
> When is this coming back?
>

Tracker never had an IPv6 record as far as I'm aware.  Nevertheless, I 
added its AAAA record to DNS.  We don't manage the reverse DNS for it, 
however, so outgoing mail to gmail addresses still fails because the 
IPv6 address lacks a PTR record.

www.ceph.com lost its IPv6 record because it is using a floating IP 
after its move to the new DreamCompute cluster.  IPv6 isn't functional 
for floating IPs in that cluster yet.

The new ceph.com host isn't using a floating IP and will have an IPv6 
record when we switch over.

Re: tracker.ceph.com

[Dan Mick]

On 12/20/2016 08:48 AM, Wido den Hollander wrote:

> I wouldn't call it a fix. In 2016 (almost 2017) IPv6 should be enabled, no questions asked.

Tell that to every single one of my internet service providers.

no, please, do tell them.  I'm tired of it too.