I've been having problems with Zope failing, resulting in a "Temporarily Unavailable" message when anyone tries to access the site. I was running Squishdot 0.7 on Zope 2.2.x and FreeBSD 3.x. Partly because I was running out of disk space, and partly because of these failures, I've just moved to a new machine with Zope 2.3.2 and FreeBSD 4.3. (In both cases, Apache's serving the http.) I was going to upgrade to the latest Squishdot as well, but ran out of time. The next morning, the failure was back, and it's happened again a week later. All I know is, the threads are still running, but nothing's accessible. If I kill the threads and restart Zope, it works again -- for a while. I haven't really grasped how all this stuff works yet, so: 1) Where to I look for clues (which log files, etc.) to help analyze the problem? 2) Is this a familiar-sounding problem? There was a reference on the Squishdot list to Apache's mod_proxy, but I'm not sure it applies to our situation... Thanks, Michael Sanborn
Michael Sanborn writes:
I've been having problems with Zope failing, resulting in a "Temporarily Unavailable" message when anyone tries to access the site. ... All I know is, the threads are still running, but nothing's accessible. If I kill the threads and restart Zope, it works again -- for a while. I have seen this problem with two lazy Zope sites (yes, many others reported similar problems). When I installed a checker process that periodically checks availability, the problem went away. I cannot explain this....
Dieter
You are referring to the -i on the command-line? Back to what you said about threads. Are you REALLY saying that I have to commit the transactions ? A word to all on this, use the detailed loggin option. Thanks to Dieter's help, I found out that Zope in my case is rear-ending dbms threads. A new user comes online and starts using the thread of the process that ran before, which basically stop Zope from responding to the world. What I want to know is why the transaction mechanism is doing this, and why didnt it ever happen in previous releases to Zope 2.3.2. Is this a bug? Paz -----Original Message----- From: zope-admin@zope.org [mailto:zope-admin@zope.org]On Behalf Of Dieter Maurer Sent: Tuesday, June 19, 2001 8:00 PM To: Michael Sanborn Cc: zope@zope.org Subject: Re: [Zope] perpetual outages Michael Sanborn writes:
I've been having problems with Zope failing, resulting in a "Temporarily Unavailable" message when anyone tries to access the site. ... All I know is, the threads are still running, but nothing's accessible. If I kill the threads and restart Zope, it works again -- for a while. I have seen this problem with two lazy Zope sites (yes, many others reported similar problems). When I installed a checker process that periodically checks availability, the problem went away. I cannot explain this....
Dieter _______________________________________________ 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 )
participants (3)
-
Dieter Maurer -
Michael Sanborn -
Paul Zwarts