RE: [Zope] Crazy Problem With Having to Restart Zope Won't Go Away!
At 04:49 PM 5/14/2001 -0500, you wrote:
have you tried including the full path to the start script?
I forgot to mention, the script resides in the Zope directory. It *does* work for a while (i.e., finds ./start like it should) then complains that it can no longer find ./start
instead of having your script sleep and then run again, why don't use just run it out of cron once every say, fifteen minutes or half hour?
I don't want the server to be down even for a minute. Yeah, maybe it's anal. Maybe a cron script *would* work, but why would it work when this infinite loop script that sleeps fails? I don't get it. Thanks for pitching in. If you have more ideas, please lay them on me. TIA, BenO
-- jacob walcik jwalcik@notwithstanding.org
On Mon, 14 May 2001, Ben Ocean wrote:
At 03:14 PM 5/14/2001 -0500, you wrote:
For some reason, from what I can judge a bug in the source code that affects some unfortunate ones such as myself, ZServer crashes every few hours and has to be re-started. So I wrote this script:
What file is this bug in? What line number?
I don't know. All I know is that from the discussion on this list last week, I discovered *several* people have had a problem with the ZServer quitting for no apparent reason since the early days of Zope. Now, that sounds like a bug to me.
So I've come up with a work-around.
You seem like you know programming and Zope better than I. Would you mind helping me? Here's my question:
For some reason, ZServer crashes every few hours and has to be re-started. So I wrote this script:
#!/bin/sh while : # if there's a page there, copy it to a test file do `wget -q -t 3 -T 30 http://thewebsons.com:8080/index_html -O test` # if there's something in the test file, go take a nap if grep '^<' test then sleep 60 # whoops! Why the heck does RH6.2 complain there's no ./start file?? Sometimes it complains, sometimes not. Hmmm... else if ./start then ./start & sleep 60 else sleep 10 fi fi done <<< and just let the fool thing run (in the Zope installation). But now I get error messages stating ./start doesn't exist! That's why the *else if* stuff, but it _still isn't good enough! Come back hours later and Zope's dead as a doornail. Furthermore, my script's not running, either. I don't get it! What am I not seeing? TIA, BenO P.S. If there were a way to keep my ZServer from crashing *without* this script, sure would like to know what it is...
_______________________________________________ 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 )
At 04:49 PM 5/14/2001 -0500, you wrote:
have you tried including the full path to the start script?
I forgot to mention, the script resides in the Zope directory. It *does* work for a while (i.e., finds ./start like it should) then complains that it can no longer find ./start
even if it's in the zope directory, the value of the environment variable $PWD should be checked to be sure it's set to the zope directory. if that script is geting thrown into the background, it should keep that value, but you never know (especially if you're getting an error reporting that the start script couldn't be found)
instead of having your script sleep and then run again, why don't use just run it out of cron once every say, fifteen minutes or half hour?
I don't want the server to be down even for a minute. Yeah, maybe it's anal. Maybe a cron script *would* work, but why would it work when this infinite loop script that sleeps fails? I don't get it. Thanks for pitching in. If you have more ideas, please lay them on me. TIA, BenO
it could run every minute out of cron... it wouldn't be perfect, but it'd keep the server up and buy you the time to find a more permanent solution.
On 14 May 2001 15:05:41 -0700, Ben Ocean wrote:
At 04:49 PM 5/14/2001 -0500, you wrote:
have you tried including the full path to the start script?
I forgot to mention, the script resides in the Zope directory. It *does* work for a while (i.e., finds ./start like it should) then complains that it can no longer find ./start
instead of having your script sleep and then run again, why don't use just run it out of cron once every say, fifteen minutes or half hour?
I don't want the server to be down even for a minute. Yeah, maybe it's anal. Maybe a cron script *would* work, but why would it work when this infinite loop script that sleeps fails? I don't get it. Thanks for pitching in. If you have more ideas, please lay them on me.
Start the server using strace, with the strace sent to a file. Then, you can look over the strace log. might be useful (strace usualy is). Perhaps even ltrace ....
participants (3)
-
Ben Ocean -
Bill Anderson -
jacob walcik