Zope does not respond but still running
Hello. I have a unknown trouble with Zope and dunno what it is. Maybe I need to post some more info? The Problem ========== Just *sometimes* browsers are waits for response forewer. Just refreshing or something else (browser restart) does not helps anymore -- I mean this is not client-side effect. Only Zope restart (via SSH) could help. And sometimes, if you wait for Zope rises up -- server does not respond anymore for ping or similar and the network card lamps are yellow and green (packet collisions? some overflow?). What Log's Sais ============ Really nothing important... What Tools Sais ============== Netstat sais, that connections are present and ports are LISTEN and "ps" sais, that Zope is running. Information ========= OS: Linux SuSE 8.0 Pro DB: Oracle Client 8i Zope: Zope 2.5.1 CPU: model name : Celeron (Coppermine) stepping : 6 cpu MHz : 797.427 cache size : 128 KB fdiv_bug : no hlt_bug : no f00f_bug : no coma_bug : no fpu : yes fpu_exception : yes cpuid level : 2 wp : yes flags : fpu vme de pse tsc msr pae mce cx8 sep mtrr pge mca cmov pat pse36 mmx fxsr sse bogomips : 1592.52 -- Regards, Bogdan Whom the gods would destroy, they first teach BASIC.
Bo M. Maryniuck wrote:
Hello.
I have a unknown trouble with Zope and dunno what it is. Maybe I need to post some more info?
The Problem ========== Just *sometimes* browsers are waits for response forewer. Just refreshing or something else (browser restart) does not helps anymore -- I mean this is not client-side effect. Only Zope restart (via SSH) could help. And sometimes, if you wait for Zope rises up -- server does not respond anymore for ping or similar and the network card lamps are yellow and green (packet collisions? some overflow?).
Just to rule out the obvious, zope's memory allocation isn't exploding and causing massive swapping? What is "top" saying in such a situation? cheers, oliver
On Thursday 07 November 2002 18:14, Oliver Bleutgen wrote:
Just to rule out the obvious, zope's memory allocation isn't exploding and causing massive swapping? No.
What is "top" saying in such a situation? Nothing special: just a little changes. Really little that no attention needed.
-- Regards, Bogdan Oh, and we just set fire to your desktop.
Hi, I have a similar problem. I am trying to reindex a lot of object (approx. 140.000) in a TextIndex and after about 10 minutes of hard work and eating up the CPU, swapping like hell and taking away 90% of the memory, the processes calm down and begin to "sleep". Just sometimes a process becomes running again for a few seconds. The swapping goes on like if he's still trying to reindex. After updating my machine (128MB RAM) to 500MB of swapspace, there's no change. The zope processes are running, but neither can I access the ZMI nor my site. Any ideas? regards Lars "Bo M. Maryniuck" schrieb:
On Thursday 07 November 2002 18:14, Oliver Bleutgen wrote:
Just to rule out the obvious, zope's memory allocation isn't exploding and causing massive swapping? No.
What is "top" saying in such a situation? Nothing special: just a little changes. Really little that no attention needed.
-- Regards, Bogdan
Oh, and we just set fire to your desktop.
_______________________________________________ Zope maillist - Zope@zope.org http://lists.zope.org/mailman/listinfo/zope ** No cross posts or HTML encoding! ** (Related lists - http://lists.zope.org/mailman/listinfo/zope-announce http://lists.zope.org/mailman/listinfo/zope-dev )
On Friday 08 November 2002 09:28, Lars H. Korte wrote:
After updating my machine (128MB RAM) to 500MB of swapspace, there's no change. The zope processes are running, but neither can I access the ZMI nor my site. Suppose some idiot will use Opera browser and clicks twenty times on the button, which does these things: in your case Zope will reindex forewer your lots of objects and in my case Zope will try rotate huge Oracle transaction.
Well, I solve this problem by some *ugly* hook, but I'm stuck to find something better. I just put/change a flag into the session, which means that method is called and in progress or already done. If user clicks some button AGAIN, Zope looks at the flag first and if it still in progress, just skips the method call. Maybe Zope gurus knows something better... In your case (might be): 1. Bad solution design? 2. Try go to the background? I mean show "Please wait" screen while indexing. But again: one user per one server?!!! If somebody ELSE over the Net will try the same thing, then Zope will index not 144.000 but 288.000 objects and your server can be exploded (or you...). 3. Try use external DB? Note, that ZODB or Gadfly is not the best choise to write something huge.
Any ideas? <wink type="veryAcid"> Use Zope only to develope empty homepage for your modest girlfriend. </wink> -- Regards, Bogdan
He who sacrifices functionality for ease of use Loses both and deserves neither
Hi, mmh? No one else but me uses this Server ;-) And I am the only one who's trying to reindex! What could it have to do with Opera? Would the change to an external DB be the only working solution? Why is the "normal way" not working with your Oracle DB? Lars "Bo M. Maryniuck" schrieb:
On Friday 08 November 2002 09:28, Lars H. Korte wrote:
After updating my machine (128MB RAM) to 500MB of swapspace, there's no change. The zope processes are running, but neither can I access the ZMI nor my site. Suppose some idiot will use Opera browser and clicks twenty times on the button, which does these things: in your case Zope will reindex forewer your lots of objects and in my case Zope will try rotate huge Oracle transaction.
Well, I solve this problem by some *ugly* hook, but I'm stuck to find something better. I just put/change a flag into the session, which means that method is called and in progress or already done. If user clicks some button AGAIN, Zope looks at the flag first and if it still in progress, just skips the method call. Maybe Zope gurus knows something better...
In your case (might be): 1. Bad solution design? 2. Try go to the background? I mean show "Please wait" screen while indexing. But again: one user per one server?!!! If somebody ELSE over the Net will try the same thing, then Zope will index not 144.000 but 288.000 objects and your server can be exploded (or you...). 3. Try use external DB? Note, that ZODB or Gadfly is not the best choise to write something huge.
Any ideas? <wink type="veryAcid"> Use Zope only to develope empty homepage for your modest girlfriend. </wink> -- Regards, Bogdan
He who sacrifices functionality for ease of use Loses both and deserves neither
On Friday 08 November 2002 16:03, Lars H. Korte wrote:
Would the change to an external DB be the only working solution? Since I dunno what you are really doing I can't answer. I just mentioned, that Gadfly or ZODB is not that great solution for biggest projects. If you want to index lots of data, IMHO you could use PostgreSQL or something.
-- Regards, Bogdan I've been in the sun for a week. I took the bold step of leaving my laptop at home. I found only 4K messages pending when I returned. -- Keith Packard
Lars H. Korte writes:
I have a similar problem. I am trying to reindex a lot of object (approx. 140.000) in a TextIndex and after about 10 minutes of hard work and eating up the CPU, swapping like hell and taking away 90% of the memory, the processes calm down and begin to "sleep". Just sometimes a process becomes running again for a few seconds. The swapping goes on like if he's still trying to reindex.
After updating my machine (128MB RAM) to 500MB of swapspace, there's no change. The zope processes are running, but neither can I access the ZMI nor my site.
Any ideas? Look at "subtransactions" in the ZCatalog "Advanced" tab. Activate them. This will reduce memory consumption.
Dieter
participants (4)
-
Bo M. Maryniuck -
Dieter Maurer -
Lars H. Korte -
Oliver Bleutgen