Re: Internal Server error

[S Ramesh]

Hi Hans,

Thanks for the mail. The www server is working fine. I tried queue-pr from
command line. It is not updating the file in the gnats-queue directory to
the respective categories. If the category name is invalid or corrupted
means, it has to move the PR to the pending directory right, but nothing is
happening. When i run queue-pr it is not prompting any errors also. I have
only one default database.

I tried with

queue-pr
queue-pr -r
queue-pr -r -f "the filename in gnats-queue directory"

nothing worked, nothing is prompted, scenario remains same. So what could
be the problem. Files in gnats-adm is fine and gnats-db.conf is also
correct. The thing is it is not prompting any errors. So no clue as such,
what could be the problem. Anybody experienced with same kind of scenario.
There is nothing found in lost+found also. Why queue-pr when run manually
with the file name also is not updating the PR to the respective category.
I tried with many categories existing, it is not updating but not giving
any errors also. Kindly help me in this regard.

Thanks.
Ramesh


Hans-Albert Schneider wrote:

> >>>>> "SR" == S Ramesh <address@hidden> writes:
>
>     SR> Internal Server Error The server encountered an internal
>     SR> error or misconfiguration and was unable to complete your
>     SR> request.
>
> This is a message from the WWW server, not from gnatsd.  Please
> check the error log of the WWW server that runs gnatsweb (or
> whatever WWW interface you use), or have its webmaster check it.
>
>     SR> Please contact the server administrator, address@hidden
>     SR> and inform them of the time the error occurred, and
>     SR> anything you might have done that may have caused the
>     SR> error.
>
> ... and also tell them to set the mail address to a value that
> also works from another machine :-).
>
>     SR> But when we create a PR, it says PR has sent
>     SR> successfully. Also it came to the gnats-queue directory.
>     SR> But when the cron runs queue-pr, the PR is not sent to
>     SR> the appropriate category.
>
> Did you try to run queue-pr manually?  Did you get an error when
> doing so?  If this is not the default database, make sure you
> specify the database directory to queue-pr.
>
>     SR> It is not moved from the gnats-queue directory. I tested
>     SR> whether cron is working or not, but it is working
>     SR> well. It happened suddenly, since it worked fine for 6
>     SR> months. What could be the problem. No settings was
>     SR> changed after installation except categories changes,
>     SR> thats it.
>
> Syntax errors in file categories, or maybe one of the other files
> in gnats-adm?
>
>     SR> But one thing happened is yesterday the server suddenly
>     SR> lost its power supply due to some loose connection and
>     SR> while restart, fsck is done. That is the only thing
>     SR> remarkably happened in the server. Is that the cause for
>     SR> the problem. If so how to fix it.
>
> A sudden power loss can cause damage at open files, so this may
> indeed be the cause.  Check the programs of your GNATS
> installation, the configuration files (gnats-db.conf as well as
> the files in gnats-adm) and the problem reports for missing or
> damaged files.  Restore them from backup, if the following hints
> do not help.
>
> If fsck finds an "inode" (internal data structure on the disk
> that represents a file) that is marked as belonging to a file but
> that has no reference from the file hierarchy, it moves that file
> to the directory "lost+found" in the same partition.  So for
> every partition of a disk, there is a directory <mount point of
> partition>/lost+found/.
>
> You find the mount points of all currently mounted filesystems by
> typing "mount". It says somthing like
>
> /proc on /proc read/write/setuid on Thu Jan 24 11:36:16 2002
> / on /dev/dsk/c0t3d0s0 [...]
> /export/home on /dev/dsk/c0t1d0s7 [...]
> /home/has on /export/home/users/has [...]
> /server/local on charlie:/export/home1/sw/server_local_sol2 [...]
>
> - /proc is not a "real" filesystem, so there is not
>   /proc/lost+found (no need for it).
>
> - /server/local is NFS mounted from another machine, so we cannot
>   tell about where lost+found for it is (the administrator of
>   machine "charlie" should know); but we need not care
>   here--"charlie" propbably did not have that power problem. :-)
>
> - /home/has is also mounted, but from local machine.  The
>   corresponding lost+found may be (and indeed is) somewhere
>   "further up".
>
> - So we have to check /lost+found and /export/home/lost+found
>   (everything that says it is "on /dev/...").
>
> The files in lost+found have some horrible names (remember that
> the relation to their place in the file hierarchy has been lost),
> so inspect everything in that directory.
>
>     SR> What is the reason or any hooks to make it work ? I think
>     SR> queue-pr is not working. But what could be the reason.
>
> The power loss may have damaged queue-pr, gnatsd, or something
> they need (dynamic library, config file, whatever).
>
> Hans-Albert
>
> --
> Hans-Albert Schneider           <address@hidden>
> Siemens AG                      phone:  (+49) 89 636 45445
> Corporate Technology            fax:    (+49) 89 636 42284
> Munich, Germany
>  -- To get my public PGP key, send me a mail with subject "send key" --