Request time grows with memory size
HI, I've encounters a performance Problem with Zope. Some requests take very long time to process while others are served very fast. All request go for the URL. The time for the delayed Requests grows with the memory size of the Zope Process. It's direct proportional. I have generated a chart showing the request times and memory size of the Zope Process. The chart can be found at http://etustar.ze.tum.de/frontend.jpg Our watchdog system restarts the Zope Process when it reaches 7.3 Gig Memory. The physical memory of the server is 16 Gig and its only used for the zope Server. There are at least 7 Gig Memory free. As you can see most of the requests are delivered within 2 seconds but some take up to 12 seconds. The Requests monitored are the request of our watchdog system. These requests are send approximately every 20 sec. When I set the restart memory limit higher the request time of the delayed request continues to grows proportionally. This server is part of a pool of 19 Servers. All are connected to one ZEO server. All show the same effect. We observe the same effect with regular requests. Some are served at normal speed but some requests are delayed. Any idea what causes this effect and how to fix it. Greetings Gerhard Schmidt -- ------------------------------------------------- Gerhard Schmidt | E-Mail: schmidt@ze.tum.de TU-München | WWW & Online Services | Tel: 089/289-25270 | Fax: 089/289-25257 | PGP-Publickey auf Anfrage
What have you done to investigate memory leaks? What external connectors are you using, like MySQL or LDAP? 2009/4/27 Gerhard Schmidt <estartu@ze.tum.de>:
HI,
I've encounters a performance Problem with Zope. Some requests take very long time to process while others are served very fast. All request go for the URL. The time for the delayed Requests grows with the memory size of the Zope Process. It's direct proportional.
I have generated a chart showing the request times and memory size of the Zope Process. The chart can be found at http://etustar.ze.tum.de/frontend.jpg
Our watchdog system restarts the Zope Process when it reaches 7.3 Gig Memory. The physical memory of the server is 16 Gig and its only used for the zope Server. There are at least 7 Gig Memory free.
As you can see most of the requests are delivered within 2 seconds but some take up to 12 seconds. The Requests monitored are the request of our watchdog system. These requests are send approximately every 20 sec.
When I set the restart memory limit higher the request time of the delayed request continues to grows proportionally.
This server is part of a pool of 19 Servers. All are connected to one ZEO server. All show the same effect.
We observe the same effect with regular requests. Some are served at normal speed but some requests are delayed.
Any idea what causes this effect and how to fix it.
Greetings Gerhard Schmidt -- ------------------------------------------------- Gerhard Schmidt | E-Mail: schmidt@ze.tum.de TU-München | WWW & Online Services | Tel: 089/289-25270 | Fax: 089/289-25257 | PGP-Publickey auf Anfrage
_______________________________________________ Zope maillist - Zope@zope.org http://mail.zope.org/mailman/listinfo/zope ** No cross posts or HTML encoding! ** (Related lists - http://mail.zope.org/mailman/listinfo/zope-announce http://mail.zope.org/mailman/listinfo/zope-dev )
-- Peter Bengtsson, work www.fry-it.com home www.peterbe.com hobby www.issuetrackerproduct.com
On Mon, Apr 27, 2009 at 12:40 PM, Peter Bengtsson <peterbe@gmail.com> wrote:
What have you done to investigate memory leaks? What external connectors are you using, like MySQL or LDAP?
It is probably not a memory leak. The graph is what I'd expect in a garbage collection scenario (ie. Python). Hedley
participants (3)
-
Gerhard Schmidt -
Hedley Roos -
Peter Bengtsson