random latency in named

[Mike Hoskins (michoski)]

FWIW as another data point we've seen the same in the wild across
RHEL/CentOS 5.x and 6.x on "large" (32 core) Xeon based servers
(E5-2650's), including 6.6 with the 2.6.32-504.16.2.el6.x86_64 kernel.
Observed while debugging other things, and haven't had time to follow up.

-----Original Message-----
From: Mathew Ian Eis <Mathew.Eis at nau.edu>
Date: Friday, May 22, 2015 at 11:33 AM
To: "bind-users at lists.isc.org" <bind-users at lists.isc.org>
Cc: Tony Finch <dot at dotat.at>
Subject: Re: random latency in named

>
>-----Original Message-----
>From: Tony Finch <dot at dotat.at>
>Date: Friday, May 22, 2015 at 2:32 AM
>To: Mathew Eis <mathew.eis at nau.edu>
>Cc: "bind-users at lists.isc.org" <bind-users at lists.isc.org>
>Subject: Re: random latency in named
>
>>Mathew Ian Eis <Mathew.Eis at nau.edu> wrote:
>>>
>>> * The OS is RHEL 6.6; we just updated the kernel to
>>> 2.6.32-504.16.2.el6.x86_64, also with no effect.
>>
>>Is your server using a Haswell CPU? If so it might be the lost futex
>>wakeup bug discussed at the links below, in which case the problem might
>>go away if you upgrade to RHEL 6.6.z.
>>
>>https://groups.google.com/forum/#!topic/mechanical-sympathy/QbmpZxp6C64
>>https://news.ycombinator.com/item?id=9542548
>
>Nope, AMD here, but that probably wouldn¹t rule it out. I think I have a
>comment somewhere on that HN thread... It looks like the futex bug
>probably affects all architectures; just some more than others, as the
>actual kernel patch references ARM.
>
>Anyhow, I wish it had been that, but the 2.6.32-504.16.2.el6.x86_64 kernel
>didn¹t fix the issue.
>
>(6.6 2.6.32-504.16.2.el6.x86_64 kernel is the 6.6.z one):
>https://rhn.redhat.com/errata/rhel-server-6.6-errata.html
>
>https://rhn.redhat.com/errata/RHSA-2015-0864.html
>
>
>Thanks,
>
>Mathew Eis
>Northern Arizona University
>Information Technology Services
>
>_______________________________________________
>Please visit https://lists.isc.org/mailman/listinfo/bind-users to
>unsubscribe from this list
>
>bind-users mailing list
>bind-users at lists.isc.org
>https://lists.isc.org/mailman/listinfo/bind-users