[Popcon-developers] Popcon takes over my PC

[Michael Bunk]

Hello Bill, hello Charles,

>> Installing anacron instead of cron would not solve your problem.  It
>> would just make sure that cronjobs that could not run when your computer
>> is switched off do run later when it is on again.
>>
>> Your problem is related to nice vs ionice, see
>> http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=483465
> 
> Do you have personal experience with this slowdown ?
> I am willing to reconsider applying the patch in 483465,
> if I am sure it fix a practical problem.

Sorry, no.  I was overly optimistic regarding my insight to this
problem.  All machines where I have popcon running are virtual machines
powerful enough for popcon to finish in a short time, leaving barely
visible traces in system monitoring (system load shown by munin for
example).  Nice and/or ionice are not necessary for them.

Anyways, since Cha(rle)s has a problem, he should really try one patch
from 483465, to see whether it helps.  I also think he has anacron
already installed, because it is unlikely his PC is running at 6:47 in
the morning...

I also wonder, why Charles said "The priority is shown as very low",
because cron does not apply any nice/ionice setting and anacron doesn't
anymore since 2.3-17 (29 Mar 2012), see
http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=373950

So if he had an anacron < 2.3-17, some nice setting would be applied by
/etc/anacrontab... and it doesn't help.

But then, from ionice(1):

"Note that before kernel 2.6.26 a process that has not asked for an I/O
priority formally  uses "none"  as  scheduling class, but the I/O
scheduler will treat such processes as if it were in the best-effort
class.  The priority within the best-effort class will be dynamically
derived from the CPU nice level of the process: io_priority = (cpu_nice
+ 20) / 5."

That means, Charles' popcon would already be running with ionice "none:
prio 4" (current versions of ionice output "unknown: prio 4", which is
another small bug, https://github.com/karelzak/util-linux/pull/46 ).

So my analysis supports Bill's resistance to 483465...

Best regards,
 Michael

-------------- next part --------------
A non-text attachment was scrubbed...
Name: smime.p7s
Type: application/pkcs7-signature
Size: 2401 bytes
Desc: S/MIME Cryptographic Signature
URL: <http://lists.alioth.debian.org/pipermail/popcon-developers/attachments/20140117/86924d27/attachment.bin>