Bugzilla Bug 177428

net-proxy/http-replicator-3.0 fails to start with actual Gentoo configuration
(i think that recent baselayout changes have break something :P)

The fix is simple: just adding the "--background" option to start section of
provided initscript (files/http-replicator-3.0.init )

See this Gentoo Forum post for further informations:
http://forums.gentoo.org/viewtopic-p-4026539.html#4026539

Reproducible: Always

Steps to Reproduce:
1. Boot system
2. rc-status shows that http-replicator is started
3. ps -ef|grep replicator doesn't show any active process about this program
4. a /etc/init.d/http-replicator restart seems to start properly it

Actual Results:  
http-replicator doesn't start properly

Expected Results:  
http-replicator should start properly

Created an attachment (id=118412) [details]
needed change to initscript

Diff of fixed initscript

*** Bug 178220 has been marked as a duplicate of this bug. ***

I've test a new installation from scratch in a Virtualbox VM.

- i've used a x86 stage3-2007.0 with a recent portage snapshot (around May,
18th)
- i've used a simple partitioning scheme (hda1-> boot , hda2-> / , hda3 ->
swap)
- i haven't updated anything, so the installed system is a "vanilla" stage3
(plus the necessary system tools and grub)
- i've installed http-replicator, added it to default runlevel and.....

IT WORKS!

:PPP

So i suppose that could be some problem in MY OWN actual Gentoo installations
:P

I'll try to update and setup the VM following my typical configurations, and i
hope to find where the things break... stay tuned 8)

Maurice, care to fix that or do you want me to fix it?

Alin, adding --background is not the right solution and I am unable to
reproduce the problem myself. 

Can you reproduce the problem? If so, we could do some debugging to find the
real cause.

(In reply to comment #5)
> Can you reproduce the problem? If so, we could do some debugging to find the
> real cause.

I can reproduce this problem on my machine.  So I can help as well.

Created an attachment (id=124346) [details]
Temporary workaround for SIGHUP

Further debugging showed that ignoring SIGHUP in the child process fixes the
problem. The solution will probably need some more changes, but as a workaround
the attached patch should do the trick.

Stay tuned.

I checked in the work-around I mentioned. If anyone notices any problems with
it, please reopen this bug.

I'm closing this bug now as RESOLVED/UPSTREAM because upstream is working on a
new release that is a rewrite and it wouldn't make much sense to spend time on
fixing the old code base if there's a work-around in place.

Thanks everyone!