Skipped content of type multipart/alternative-------------- next part -------------- A non-text attachment was scrubbed... Name: smime.p7s Type: application/x-pkcs7-signature Size: 4490 bytes Desc: not available Url : http://modevia.com/pipermail/py-transports/attachments/20060615/d6768747/smime.bin From [EMAIL PROTECTED] Thu Jun 15 00:14:01 2006 From: [EMAIL PROTECTED] (James Bunton) Date: Thu Jun 15 09:09:54 2006 Subject: [py-transports] very high load using pymsnt latest SVN for ~30s In-Reply-To: <[EMAIL PROTECTED]> References: <[EMAIL PROTECTED]> <[EMAIL PROTECTED]> <[EMAIL PROTECTED]> <[EMAIL PROTECTED]> Message-ID: <[EMAIL PROTECTED]>
On 09/06/2006, at 4:29 AM, Lucas Nussbaum wrote: > On 08/06/06 at 16:06 +0200, Hylke van der Schaaf wrote: >> James Bunton wrote: >>> On 01/06/2006, at 9:21 PM, Lucas Nussbaum wrote: >>> >>>> Hi, >>>> >>>>> From time to time, pymsnt starts using a lot of CPU for about >>>>> 30s, while >>>> it doesn't use that much the rest of the time. No message is logged >>>> during this time period. >>>> >>>> Those high load periods can be seen on >>>> http://apijab.apinc.org/munin/apinc.org/apijab.apinc.org-cpu.html >>>> pymsnt is nearly the only thing running on this system, and I've >>>> verified that pymsnt was the cause for the high CPU usage. >>>> >>>> Sometimes, it's longer or shorter that 35s : >>>> 13:08:38 -> 13:09:17 >>>> >>>> >>>> What does QNG mean ? >>>> What could be the source for this ? >>>> I'm using pymsnt r168. >>>> >>>> I'm available for testing/debugging if needed. >>> >>> >>> :| >>> >>> That's not good. Has it just started doing it recently? >>> QNG is just a ping. It shouldn't have anything to do with it. >>> >>> Garbage collection maybe? >>> >>> Anybody got any ideas? >> >> I would guess garbage collection. >> How is the memory usage of python when this happens? How much >> memory is >> in the system? how much is used by python, how much for buffers, how >> much swap is used? >> >> If python has to garbage collect on things that where swapped out >> this >> can cause a high load. > > I didn't have python's memory usage specifically, but there was a > lot of > memory available, and no swap used. Maybe python starts garbage > collecting even if there's still some memory available ? > > Anyway, it hasn't done it again since then, so I'm not sure ... If anybody sees this again, or can find steps to reproduce it, or has any clues what caused it, please let me know. Otherwise we'll just hope it was a freak occurrence :) --- James
