[hobbit] Critical Systems view loading problem

[Gary Baluha]

As I previously posted, I get this problem every now and then as well.
About a month back, the Critical Systems page suddenly become useless when
it became stuck with that "Internal Server Error" issue.  My co-worker came
across an apparent fix that the file permissions for the
hobbit-nkview.cfgfile were wrong, and the --debug option in
hobbitcgi.cfg for hobbit-nkview.cfg was preventing the page from loading.
This now appears NOT to be the case, because the eternal Internal Server
Error problem is back.  It seems it was just coincidence that he made the
changes when the Critical Systems page started working again.

Also, while I was in the process of typing the above section, it appears the
Critical Systems page is working again.  I made absolutely no changes to
anything during this time.  Unfortunately now, as before, I cannot determine
any causal relationship.  Additionally, unlike Tracy's problem below, it
doesn't appear to be related to the alerts that are showing up either (I can
confirm that no alert statuses changed while I was writing this).

I'm going to have to go with Tracy's assesment that it is a pointer issue as
pointed out.  I do recall during my programming days of incorrect pointer
usage in the code causing intermittent and non-reproducible errors
occuring...  Unfortunately, it's been a while since I've programmed in
C/C++, and I would have to spend a while with the code to see if this really
is the issue, and how to fix it.  All I know is, it sounds plausible.

Anyone else have any ideas, or am I just going a little off the deep end
with this (which is quite possible)?


On Sep 17, 2007 4:43 PM, Tracy Di Marco White <gendalia at gmail.com> wrote:

> On 9/7/07, Henrik Stoerner <henrik at hswn.dk> wrote:
> > On Thu, Sep 06, 2007 at 09:30:58PM -0500, Tracy Di Marco White wrote:
> > > I'm getting an "Internal Server Error" and the error log shows
> > > "Premature end of script headers: hobbit-nkview.sh".  My problem seems
> > > to be related to a test being yellow right now, and right now being
> > > outside of the parameters of when the machine/test combo is critical.
> > > If I change the critical time for the event from "|W:0800:1700|" to
> > > "||", the critical systems page comes up fine.  If I put the time
> > > constraints back, the page fails to come up again.  It started failing
> > > after 1700, although I didn't notice it for about 15 minutes.  Is
> > > anyone else seeing this problem?
> >
> > Interesting, it does sound like a bug. Could you send me that line from
> > the hobbit-nkview.cfg file ?
>
> So I worked with a co-worker to figure out where the problem is.
>  In get_nkconfig() in ./lib/loadnkconf.c
>
>                        /* Go to the next */
>                        handle = rbtNext(rbconf, handle);
>                        if (handle != rbtEnd(rbconf)) {
>                                rbtKeyValue(rbconf, handle, &k1, &k2);
>                                if (strncmp(realkey, ((nkconf_t
> *)k2)->key,...
>
>  here k2 is treated as a pointer to a nkconf_t
>  but it appears that it is really a 'char *'.
>
> Is what he concluded from our poking with gdb. We haven't gotten
> any further in debugging, but hoped this would help to pin down
> the problem.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.xymon.com/pipermail/xymon/attachments/20071224/b40a7989/attachment.html>