Questions[zope](server)

Hi I am running multiple applications on the same machine like Plone which is using ZOPE application server that is running on port 8080 and Tomcat which is hosting multiple web applications and running on port 8081.

I want to redirect all my incoming traffic from apache server which is running on port 80 to these server, so that external world will see only apache's default port 80. I tried many configurations for doing the same but none of them seems to be working for both the severs but yes individually if I am configuring its working flawlessly. So, only one configuration is getting used and the other has no effect.

My configuration looks like: Pastebin Link

I also tried with the following configuration:

 <VirtualHost *:80>
        ServerName web.url.in
        ProxyPreserveHost On
        ProxyPass /MNCD2016 http://web.url.in:8081/MNCD2016
        ProxyPassReverse /MNCD2016 http://web.url.in:8081/MNCD2016
        RewriteEngine on
        RewriteRule ^/$ http://web.url.in/MNCD2016
    </VirtualHost>

    <VirtualHost *:80>
        ServerName web.url.in
        ProxyPreserveHost On
        ProxyPass / http://web.url.in:8080/
        ProxyPassReverse / http://web.url.in:8080/
        RewriteEngine on
        RewriteRule ^/$ http://web.url.in/ckpw
    </VirtualHost>

The problem is that the configuration works only for either one of the port not for both. For example if the Plone rewrite rule is above it will take the precedence and the other tomcat configuration will not work. Individually they work smoothly.

How can I make them work simultaneously for both the servers (Zope, Tomcat) may be even more in future. So, that I can divert all incoming request from apache to these servers and the external world should see only apache in the front not these servers.

This is the link which I followed for Plone configuration.

I keep running into this issue when I need to temporarily end a Zope server in order to do something. If I run the script

./bin/zopectrl fg or ./bin/isntance fg

It launches zope's server gets the ports it needs and starts serving. The only way I found to end this command is to do ctrl-z and then disown %1 however once I start it up, the resources it took the first time are blocked from being used again. THe only way I know to fix this is to restart the whole server which is not efficient.

How to I stop the script properly so I can start it up again?

We're running Plone 4.1, Zope 2.13 behind Apache with mod_proxy_balancer. We have 2 zope instances running 2 threads each.

We're seeing a number of connections between Zope and Apache being ESTABLISHED but never released (or at least they're hanging around a long time).

This is happening on our live and test system. Trying to analyse this on our test system:

pid of one of the zope instances is 17577

# netstat -a -p | fgrep 17577/
tcp        0      0 *:18091                     *:*                         LISTEN      17577/python
tcp        0      0 dev.ic.soschildrensvi:18091 dev.ic.soschildrensvi:36932 ESTABLISHED 17577/python
tcp        0      0 dev.ic.soschildrensvi:18091 dev.ic.soschildrensvi:36931 ESTABLISHED 17577/python
tcp        0      0 dev.ic.soschildrensvi:18091 dev.ic.soschildrensvi:36929 ESTABLISHED 17577/python
tcp        0      0 dev.ic.soschildrensvi:18091 dev.ic.soschildrensvi:36940 ESTABLISHED 17577/python
tcp        0      0 dev.ic.soschildrensvi:18091 dev.ic.soschildrensvi:36936 ESTABLISHED 17577/python
tcp        0      0 localhost.localdomain:51672 localhost.localdomain:mysql ESTABLISHED 17577/python
tcp        0      0 localhost.localdomain:47304 localhost.localdomain:mysql ESTABLISHED 17577/python
tcp        0      0 localhost.localdomain:47297 localhost.localdomain:mysql ESTABLISHED 17577/python
tcp        0      0 dev.ic.soschildrensvi:18091 dev.ic.soschildrensvi:47559 ESTABLISHED 17577/python
tcp        0      0 dev.ic.soschildrensvi:18091 dev.ic.soschildrensvi:47556 ESTABLISHED 17577/python
tcp        0      0 dev.ic.soschildrensvi:18091 dev.ic.soschildrensvi:47554 ESTABLISHED 17577/python
tcp        0      0 dev.ic.soschildrensvi:18091 dev.ic.soschildrensvi:47552 ESTABLISHED 17577/python
tcp        0      0 dev.ic.soschildrensvi:18091 dev.ic.soschildrensvi:47567 ESTABLISHED 17577/python
tcp        0      0 dev.ic.soschildrensvi:18091 dev.ic.soschildrensvi:47565 ESTABLISHED 17577/python
tcp        0      0 dev.ic.soschildrensvi:18091 dev.ic.soschildrensvi:47560 ESTABLISHED 17577/python

So 13 connections between Apache and Zope. I make a request to Plone and this increases 16 then 17 then 19. It was up to 88 on one instance and 89 on the other before I restarted them this morning. Apache configuration looks like this

ProxyPreserveHost On
<Proxy balancer://zopes>
  BalancerMember http://dev:18091
  BalancerMember http://dev:18092
</Proxy>
RewriteRule ^/(.*) balancer://zopes/VirtualHostBase/http/%{SERVER_NAME}:%{SERVER_PORT}%{ENV:PLONE_FOLDER_PATH}/VirtualHostRoot/$1 [last,proxy]

The reason I'm posting this is because we're experiencing some connectivity and performance issues on our servers and I'm wondering whether this might be something to do with it.

We also have a separate Plone 3.3 / Zope 2.10 / Apache with no balancer configuration and this doesn't leave any connections lying around.

I'm just making some changes to the way we run Zope following a recent security advisory. In the advisory it says:

“Make sure that the Zope/Plone service is running with with minimum privileges. Ideally, the Zope and ZEO services should be able to write only to log and data directories.”

Currently we have a dedicated linux user that runs buildout, starts and stops the server and which runs the Zope service. We're using supervisord to manage the two Zope instances we run in parallel (relstorage backend).

Our current zope user needs write permission on more than just the log and data directory so that it can run buildout so I opted to create an additional linux user, zoperun, to actually run the zope service. zoperun is in the same group as zope and only has write permissions on the var directory. This is all working fine except one thing: I'd like the original zope user to be able to start and stop the zope instances using supervisorctl.

I can't get this working. For example, when I try to run bin/supervisorctl status I get

error: , [Errno 13] Permission denied: file: line: 1

I think this is because the supervisor socket doesn't have group read and write permissions

srw------- 1 zoperun zopeour 0 Sep 30 09:00 supervisor.sock

I've set umask = 002 in the supervisord configuration but this has no affect on this superctl socket.

Is there any way I can achieve what I'd like?

Afternoon, I'm running Plone 2.5.3 on SLES 10. With a separate Apache server (192.168.6.80) using mod rewrite to access Plone.

Recently been experiencing slow performance loading pages. When I checked the ZOPE events.log there are lots of:

2011-02-09T13:20:18 ERROR ZServer uncaptured python exception, closing channel <ZServer.HTTPServer.zhttp_channel connected 192.168.6.80:42815 at 0x2aaaacaf97a0 channel#: 349 requests:> (socket.error:(104, 'Connection reset by peer') [/opt/Plone-2.5.3/Python-2.4.4/lib/python2.4/asynchat.py|initiate_send|219] [/opt/Plone-2.5.3/lib/python/ZServer/medusa/http_server.py|send|417] [/opt/Plone-2.5.3/Python-2.4.4/lib/python2.4/asyncore.py|send|332])

I found a post which suggested this error is just caused by users navigaing away from a page before the request is complete, can anyone confirm this is normal Zope behaviour? Maybe I just need to reduce the log level?

Cheers,

Nathan.