0000011: mod_caucho loses correct configuration knowledge

11

minor

always

03-01-05 00:00

05-11-05 00:00

Reporter:

ferg

Platform:

Assigned To:

OS:

Priority:

high

OS Version:

Status:

closed

Product Version:

Product Build:

Resolution:

fixed

Projection:

none

ETA:

none

Fixed in Version:

3.0.13

Summary:

0000011: mod_caucho loses correct configuration knowledge

Description:

RSN-1
(rep by M Moriyama)

If Resin runs with mod_caucho + Apache 2 (in this state, everything is working ok, e.g. eos.wazap.jp/index.jsp is served correcly) and:

Resin is stopped (while Apache is running),
a page which would normally be servered by Resin (e.g. eos.wazap.jp/index.jsp) is accessed
Resin is started

mod_caucho seems to forget all settings and keeps on returning 503s for all jsps.

The attached file "caucho-status-working" shows caucho-status after a fresh "Resin->Apache" start, everything is working. The file "caucho-status-not-working" shows the caucho-status just after Resin has been started again (no. 3. above); if the same page (e.g. eos.wazap.jp/index.jsp) is accessed, only a 503 is shown.

With this state, an Apache stop -> start doesn't help, caucho-status shows the same status "caucho-status-not-working". Neither does a Resin stop, Apache stop, Resin start, Apache start help, caucho-status shows still the not working status.

What only helps in this situation is to delete the cached config file /tmp/192.168.10.14_6802 and to restart Apache. The file "192.168.10.14_6802.working" is the cached config file while mod_caucho is working ok. "192.168.10.14_6802.not_working" is the file after mod_caucho has forgotten all settings and which I have to delete.

While caucho-status shows the "not-working-status", Resin itself, which we access via port 9090, serves the jsps without any problems.

We see this on Red Hat 9, JDK 1.4.2, Apache 2.0.52. 192.168.10.14 and 202.218.5.43 are the same machine ("eos").

In httpd.conf, we have:
<VirtualHost 202.218.5.43:80>
   ServerName eos.wazap.jp
   ResinConfigServer 192.168.10.14 6802
   AddHandler caucho-request .jsp .wz
   <Location "/*.jsp">
       Order deny,allow
       Allow from all
       SetHandler caucho-request
   </Location>
   <Location "/*.wz">
       Order deny,allow
       Allow from all
       SetHandler caucho-request
   </Location>

   <Location "/eastbeam">
       SetHandler caucho-request
   </Location>

   CauchoStatus true
   <Location /eastbeam/intern/admin/caucho-status>
       SetHandler caucho-status
   </Location>

   <Location /WEB-INF/*>
       Deny from all
   </Location>
   <Location /lib/*>
       Deny from all
   </Location>
...
</VirtualHost>

In resin.conf, we have:
       <http server-id="eos" host="202.218.5.43" port="9090"/>
       <cluster>
           <srun server-id="eos" host="192.168.10.14" port="6802"
index="2"/>
       </cluster>
...
       <host id='wazap.jp'>
           <host-alias>eos.wazap.jp</host-alias>
...
       </host>

Steps To Reproduce:

Additional Information:

Relationships

Attached Files:

Mantis - Resin
Viewing Issue Advanced Details

ID:	Category:	Severity:	Reproducibility:	Date Submitted:	Last Update:
11		minor	always	03-01-05 00:00	05-11-05 00:00

Reporter:	ferg	Platform:
Assigned To:		OS:
Priority:	high	OS Version:
Status:	closed	Product Version:
Product Build:		Resolution:	fixed
Projection:	none
ETA:	none	Fixed in Version:	3.0.13

Summary:	0000011: mod_caucho loses correct configuration knowledge
Description:	RSN-1 (rep by M Moriyama) If Resin runs with mod_caucho + Apache 2 (in this state, everything is working ok, e.g. eos.wazap.jp/index.jsp is served correcly) and: Resin is stopped (while Apache is running), a page which would normally be servered by Resin (e.g. eos.wazap.jp/index.jsp) is accessed Resin is started mod_caucho seems to forget all settings and keeps on returning 503s for all jsps. The attached file "caucho-status-working" shows caucho-status after a fresh "Resin->Apache" start, everything is working. The file "caucho-status-not-working" shows the caucho-status just after Resin has been started again (no. 3. above); if the same page (e.g. eos.wazap.jp/index.jsp) is accessed, only a 503 is shown. With this state, an Apache stop -> start doesn't help, caucho-status shows the same status "caucho-status-not-working". Neither does a Resin stop, Apache stop, Resin start, Apache start help, caucho-status shows still the not working status. What only helps in this situation is to delete the cached config file /tmp/192.168.10.14_6802 and to restart Apache. The file "192.168.10.14_6802.working" is the cached config file while mod_caucho is working ok. "192.168.10.14_6802.not_working" is the file after mod_caucho has forgotten all settings and which I have to delete. While caucho-status shows the "not-working-status", Resin itself, which we access via port 9090, serves the jsps without any problems. We see this on Red Hat 9, JDK 1.4.2, Apache 2.0.52. 192.168.10.14 and 202.218.5.43 are the same machine ("eos"). In httpd.conf, we have: <VirtualHost 202.218.5.43:80> ServerName eos.wazap.jp ResinConfigServer 192.168.10.14 6802 AddHandler caucho-request .jsp .wz <Location "/.jsp"> Order deny,allow Allow from all SetHandler caucho-request </Location> <Location "/.wz"> Order deny,allow Allow from all SetHandler caucho-request </Location> <Location "/eastbeam"> SetHandler caucho-request </Location> CauchoStatus true <Location /eastbeam/intern/admin/caucho-status> SetHandler caucho-status </Location> <Location /WEB-INF/> Deny from all </Location> <Location /lib/> Deny from all </Location> ... </VirtualHost> In resin.conf, we have: <http server-id="eos" host="202.218.5.43" port="9090"/> <cluster> <srun server-id="eos" host="192.168.10.14" port="6802" index="2"/> </cluster> ... <host id='wazap.jp'> <host-alias>eos.wazap.jp</host-alias> ... </host>
Steps To Reproduce:
Additional Information:
Relationships
Attached Files: