Bug 281704 – net-proxy/http-replicator-3.0-r2 does not start anymore

• Product: Gentoo Linux
• Status: RESOLVED DUPLICATE

[Description Stefan Borschtel 2009-08-16 19:51:47 UTC]

When starting http-replicator-3.0-r2 via the init-script (/etc/init.d/http-replicator start) the programm does not start anymore. No error-messages are printed in the logs i found. So i searched for a solutions at forums.gentoo.org. There someone told me that he is suffering the same problem after python 2.6 was marked stable. So maybe it is related to a false exit-code.
 When launched manually the programm works nicely. I wrote myself a dirty work-around by using a script:

#!/bin/bash
start-stop-daemon --start --pidfile /var/run/http-replicator.pid --name http-replicator --exec /usr/bin/http-replicator -- -s -f --pid /var/run/http-replicator.pid --daemon --dir /var/cache/http-replicator --user portage --alias /usr/portage/packages/All:All --log /var/log/http-replicator.log --ip 192.168.*.* --ip 10.*.*.* --port 8080

This works fine. So either the exit-codes of http-replicator or the initd-scrtipts needs reviewing.

Here is a link to the forum-thread:

http://forums.gentoo.org/viewtopic-t-787761-highlight-httpreplicator.html


Reproducible: Always

Steps to Reproduce:
1.run /etc/init.d/http-replicator start
2.
3.

Actual Results:  
Does not start, but there are no error-messages.

Expected Results:  
It starts by using "/etc/init.d/http-replicator start"

[Comment 1 ??? 2009-08-17 08:33:03 UTC]

Same for me here.

Works fine when running the command manually - it just seems something with the init script is broken...

[Comment 2 Maurice van der Pot (RETIRED) 2009-08-18 16:45:47 UTC]

I am unable to reproduce this problem. Maybe it is related to bug #177428.

Does the init script always fail for you guys? Even if you run it many times, as with: for i in `seq 1 30`; do /etc/init.d/http-replicator restart; done

In order to debug this, I'd need to get trace of what happens when it fails.
Could one of you put "strace -f" in front of start-stop-daemon in the init script and "&> /tmp/strace.log" after $DAEMON_OPTS?

If you then start http-replicator the init script will not show the correct output status, but you'll have a log in /tmp/strace.log that should tell us more.

[Comment 3 ??? 2009-08-19 10:13:49 UTC]

Created attachment 201700 [details]
Added "strace" (without -f) to the init script

[Comment 4 ??? 2009-08-19 10:16:16 UTC]

Created attachment 201701 [details]
Added "strace -f" to the init script - waited until http-replicator got started, then hit Ctrl-C

[Comment 5 ??? 2009-08-19 10:17:06 UTC]

Ok, I tried running the init script multiple times, but it doesn't make any difference...

Then I tried adding "strace -f" to the init script which resulted in the trace process not returning as it is waiting for the daemon process to exit which obviously won't exit (but is started correctly)...
Then I added only "strace" to the init script. Now when running the init script http-replicator is started properly and the init script doesn't detect any error and says "ok" (ok, that "ok" probably comes from strace exiting successfully...).

Anyway, I attached logs for both cases...

[Comment 6 Stefan Borschtel 2009-08-22 18:20:13 UTC]

Same for me: I used the above FOR-loop to start http-replicator 30 times in a row and it always shows the same output:

 * Starting Http-Replicator ...
 * Failed to start Http-Replicator                                        [ !! ]

When i add the above mentioned changes to the init-script, then the script does not terminate. I have to press CTRL+C to stop it. I will attach the output.

[Comment 7 Stefan Borschtel 2009-08-22 18:21:19 UTC]

Created attachment 201952 [details]
My 'strace -f'-output received by modifying the script as mentioned.

[Comment 8 Bill Kenworthy 2009-08-29 05:42:10 UTC]

Ran into this as well after the python-2.6 upgrade.  http-replicator is seeming to return non-zero on start now.  Adding --oknodo to the start-stop-daemon in the start stanza fixed it.

BillK

[Comment 9 Stefan Borschtel 2009-08-29 10:06:48 UTC]

I already wrote it on the GentooForum,but ... thank you again, Bill.
Seems to work.

[Comment 10 Martin Ereth 2009-10-02 08:12:46 UTC]

This bug exists here too. (gentoo x86)
I've tried adding --oknodo. It shows ok but it doesn't work.
Maybe anybody is able to update the ebuild?
Thanks a lot!

[Comment 11 Vadim Efimov 2009-10-05 09:11:09 UTC]

may be it related to
http://archives.gentoo.org/gentoo-dev/msg_54ec9c3d4c15c3f52e4c71fef5d42633.xml

(In reply to comment #10)
> This bug exists here too. (gentoo x86)
> I've tried adding --oknodo. It shows ok but it doesn't work.
> Maybe anybody is able to update the ebuild?
> Thanks a lot!
> 
 works fine for me now, on amd64:
-start-stop-daemon --start --pidfile /var/run/http-replicator.pid --name http-replicator \
+start-stop-daemon --start --pidfile /var/run/http-replicator.pid --name http-replicator --oknodo \

[Comment 12 Kelly Price 2009-10-17 14:16:49 UTC]

I also have that here on a AMD64 running system.  Using --oknogo in the Init.d script works.

[Comment 13 Maurice van der Pot (RETIRED) 2009-10-19 19:15:03 UTC]

It's actually not http-replicator's fault apparently.

*** This bug has been marked as a duplicate of bug 286191 ***

[Comment 14 Maurice van der Pot (RETIRED) 2009-10-19 19:17:16 UTC]

Don't forget to subscribe to the dupe to get updates on this issue.