<div dir="ltr">Indefinite wait buffer typically means the disk driver locked up while the system was trying to swap to/from it. Two questions: (1) Is the swap encrypted ? And (2) What is the device attachment for the swap drive (in the dmesg output).<div><br></div><div>-Matt</div></div><div class="gmail_extra"><br><div class="gmail_quote">On Mon, Aug 8, 2016 at 5:18 AM, Stefan Unterweger <span dir="ltr"><<a href="mailto:232.20711@chiffre.aleturo.com" target="_blank">232.20711@chiffre.aleturo.com</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">Und täglich grüßt das Murmeltier… :o/<br>
<br>
<br>
Another crash with 09b0c4f76, but this one looks different. The console<br>
was scrolling endless streams of lines like this one:<br>
<br>
| swap_pager: indefinite wait buffer: bp 0xffffffe00e57bea8 offset: 4133535744, size: 4096<br>
<br>
With varying values for both ‚bp‘ and ‚offset‘. No kernel panic,<br>
though.<br>
<br>
The system was swapping when it happened; I was toying with the dports<br>
git repository. I have seen a crash like this once before, but this was<br>
months ago. Is this related, or something new altogether?<br>
<br>
Reboot went without any problems, so -that- problem really does seem to<br>
be resolved.<br>
<br>
Stefan<br>
<br>
<br>
* Stefan Unterweger on Sun, Aug 07, 2016 at 02:49:31AM +0200:<br>
> I stalled the update a bit because I wanted to make sure that the boot<br>
> problems were resolved first (our customers are already edgy; another<br>
> two-hour downtime just for booting would have been too much).<br>
><br>
> Given that 4.6 is now out, I’ve decided to take the plunge and do it.<br>
> Booting went without a fault (could be coincidence; I will see if the<br>
> problem is gone next time when I’ll re-add more storage), initial quick<br>
> stress tests went rather well. As far as I can tell, I/O performance<br>
> has approximately doubled.<br>
><br>
> I have scheduled a few heavy stress tests for the upcoming weekend where<br>
> I will try to torture the system with everything I got, then I’ll report<br>
> back.<br>
><br>
> Once again: Thanks BIG time for all the wonderful work that all of you<br>
> have put into Dragonfly, I really enjoy working with it.<br>
><br>
> Cheers,<br>
> Stefan<br>
<br>
> * Matthew Dillon on Tue, Jul 26, 2016 at 06:03:03PM -0700:<br>
> > On this issue, we've made a bunch of fixes in just the last few days, you<br>
> > will definitely want to update to the latest master. Things are looking a<br>
> > lot more stable now. We are continuing to life test three server<br>
> > configurations this week to locate any remaining low-hanging fruit.<br>
> ><br>
> > -Matt<br>
<span class="HOEnZb"><font color="#888888"><br>
--<br>
▪ Die Internetbleibe. Schick, magisch, leistungsstark. <a href="https://internetbleibe.de/" rel="noreferrer" target="_blank">https://internetbleibe.de/</a><br>
▪ medoly media UG (haftungsbeschränkt) | Hausburgstraße 13, 10249 Berlin<br>
▪ <a href="mailto:info@medolymedia.de">info@medolymedia.de</a> | <a href="https://medolymedia.de/" rel="noreferrer" target="_blank">https://medolymedia.de/</a> | Tel. 030 609 826‒560 | Fax …‒569<br>
▪ Geschäftsführer: Matthias Nothhaft | HRB 131198 (Amtsgericht Berlin-Charlottenburg), Sitz: Berlin, USt-ID: DE275221203<br>
</font></span></blockquote></div><br></div>