Re: AssertException in process_common.c [resurrecting old thread]

[Martin Pala]

This probably explains the problem :) These extensions such as VServer 
or LIDS, etc. are able to hide processes.

Monit builds the process tree for analysis starting from its root, which 
is process 1 (init). Because this process is on your system almost 
certainly hidden (but based on PPID value of 'orphaned' processes it 
exists), monit will fail to build the tree.

I think that it will be better to unhide init process in all contexts 
then do some workaround in monit. The reason is, that init process (PID 
= 1) is well-known and even as you can see, you can't hide it completely 
(see the orphans PPID value). The 'unkillable' protection is probably 
more adequate.

Martin

Allen Shaw wrote:
> [replying to list....]
>
> Hi Christian,
>
> Yeah, you're right, there's no init process in that list, is there.  
> Ha.  Okay, maybe it's relevant that this is a "virtual dedicated server" 
> environment.  Gee, I should have said that before!  Just didn't think of 
> it at all.  That would explain the other "lot of missing processes" 
> wouldn't it.
>
> Um, so does this make problems for monit, and is there a way to work 
> around those?
>
> As for the kernel, I don't know exactly what's in there, because that 
> level of administration is done by the VDS provider, not me.  And if UML 
> is User Mode Linux, then I'm pretty sure the anser is "no."
>
> Thanks,
> Allen
>
> Christian Hopp wrote:
>
> > On Mon, 10 Jan 2005, Allen Shaw wrote:
> >
> > Hi Allen!
> >
> > Mhhh... are you using any special security patches in your kernel or 
> > are you running in UML?  Because monit is right.  There is no process 
> > with pid=1 which is in 99.9999% of all cases the init process.
> > I am missing a lot of other processes in your "ps -ef".  Strange!
> >
> > Christian
> >
> > > I'm including here the process list to match with monit debug output
> > > from the command `monit -v -I` with monit version 4.5-cvs20050106
> > > (posted to the list, and copied here for reference).
> > >
> > > Hope this is helpful.
> > >
> > > Thanks,
> > > Allen
> > >
> > > =======================
> > > process list:
> >
> >
> > (...)