<div dir="ltr"><div dir="ltr"><br></div><br><div class="gmail_quote"><div dir="ltr" class="gmail_attr">On Fri, Sep 4, 2020 at 11:12 PM Tim Burke <<a href="mailto:tburke@nvidia.com">tburke@nvidia.com</a>> wrote:<br></div><blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex">On 9/3/20 3:30 AM, Dmitry Tantsur wrote:<br>
> *External email: Use caution opening links or attachments*<br>
> <br>
> <br>
> Hi folks,<br>
> <br>
> I'm trying to revive the Bifrost stable/stein CI, and after fixing a <br>
> bunch of issues in <a href="https://review.opendev.org/749014" rel="noreferrer" target="_blank">https://review.opendev.org/749014</a> I've hit a wall <br>
> with what seems an eventlet problem: ironic-inspector fails to start with:<br>
> <br>
> Exception AttributeError: "'_SocketDuckForFd' object has no attribute <br>
> '_closed'" in <bound method _SocketDuckForFd.__del__ of <br>
> _SocketDuckForFd:16> ignored<br>
> <br>
> I've managed to find similar issues, but they should have been resolved <br>
> in the eventlet version in stein (0.24.1). Any ideas?<br>
> <br>
> If we cannot fix it, we'll have to EOL stein and earlier on bifrost.<br>
> <br>
> Dmitry<br>
> <br>
> -- <br>
> Red Hat GmbH, <a href="https://de.redhat.com/" rel="noreferrer" target="_blank">https://de.redhat.com/</a> , Registered seat: Grasbrunn,<br>
> Commercial register: Amtsgericht Muenchen, HRB 153243,<br>
> Managing Directors: Charles Cachera, Brian Klemm, Laurie Krebs, Michael <br>
> O'Neill<br>
<br>
The "ignored" makes me think that it shouldn't actually be a problem -- <br>
are we assuming that's the error because of logs like <br>
<a href="https://c972f4bb262ae2d5c5d6-598e1d61c0aab85aa3b67b337ca2c556.ssl.cf2.rackcdn.com/749014/2/check/bifrost-integration-tinyipa-ubuntu-xenial/9d2905c/logs/ironic-inspector.log" rel="noreferrer" target="_blank">https://c972f4bb262ae2d5c5d6-598e1d61c0aab85aa3b67b337ca2c556.ssl.cf2.rackcdn.com/749014/2/check/bifrost-integration-tinyipa-ubuntu-xenial/9d2905c/logs/ironic-inspector.log</a> <br>
?<br>
<br>
Digging down to <br>
<a href="https://c972f4bb262ae2d5c5d6-598e1d61c0aab85aa3b67b337ca2c556.ssl.cf2.rackcdn.com/749014/2/check/bifrost-integration-tinyipa-ubuntu-xenial/9d2905c/logs/all/syslog" rel="noreferrer" target="_blank">https://c972f4bb262ae2d5c5d6-598e1d61c0aab85aa3b67b337ca2c556.ssl.cf2.rackcdn.com/749014/2/check/bifrost-integration-tinyipa-ubuntu-xenial/9d2905c/logs/all/syslog</a> <br>
shows tracebacks like<br>
<br>
File ".../eventlet/hubs/__init__.py", line 39, in get_default_hub<br>
import eventlet.hubs.epolls<br>
File ".../eventlet/hubs/epolls.py", line 13, in <module><br>
from eventlet.hubs.hub import BaseHub<br>
File ".../eventlet/hubs/hub.py", line 24, in <module><br>
import monotonic<br>
File ".../monotonic.py", line 169, in <module><br>
raise RuntimeError('no suitable implementation for this system: ' + <br>
repr(e))<br>
RuntimeError: no suitable implementation for this system: <br>
AttributeError("'module' object has no attribute 'epolls'",)<br>
<br>
Maybe it's worth looking at why monotonic can't find a suitable <br>
implementation?<br></blockquote><div><br></div><div>It used to be an issue in eventlet, but we seem to be using a new enough version to avoid it (the initial problem was IIRC around the pike timeframe).</div><div><br></div><div>Dmitry<br></div><div> </div><blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex">
<br>
Tim<br>
<br>
</blockquote></div><br clear="all"><br>-- <br><div dir="ltr" class="gmail_signature"><div dir="ltr">Red Hat GmbH, <a href="https://de.redhat.com/" target="_blank">https://de.redhat.com/</a> , Registered seat: Grasbrunn, <br>Commercial register: Amtsgericht Muenchen, HRB 153243,<br>Managing Directors: Charles Cachera, Brian Klemm, Laurie Krebs, Michael O'Neill <br></div></div></div>