[Spacewalk-list] SW 1.9 abrt crash upload fails

Milan Zazrivec mzazrivec at redhat.com
Fri Mar 8 17:19:31 UTC 2013 

> Hello,
> For the duplicate dir, the problem is that, when a duplicate dir is
> found... no data is uploaded.
> To test sosreport file upload, I've modified the configuration and reran my
> test using my simple loop.
> I get these logs:
> Mar  8 15:47:29  abrtd: Directory 'ccpp-2013-03-08-15:47:29-20599' creation
> detected
> Mar  8 15:47:29  abrt[20607]: Saved core dump of pid 20599 (/bin/bash) to
> /var/spool/abrt/ccpp-2013-03-08-15:47:29-20599 (413696 bytes)
> Mar  8 15:47:38  abrtd: Sending an email...
> Mar  8 15:47:38  abrtd: Email was sent to: root at localhost
> Mar  8 15:47:38 abrtd: New problem directory
> /var/spool/abrt/ccpp-2013-03-08-15:47:29-20599, processing
> Mar  8 17:05:23 abrt[21793]: Saved core dump of pid 21781 (/bin/bash) to
> /var/spool/abrt/ccpp-2013-03-08-17:05:23-21781 (413696 bytes)
> Mar  8 17:05:23  abrtd: Directory 'ccpp-2013-03-08-17:05:23-21781' creation
> detected
> Mar  8 17:05:32  abrtd: Sending an email...
> Mar  8 17:05:32  abrtd: Email was sent to: root at localhost
> Mar  8 17:05:32  abrtd: Duplicate: UUID
> Mar  8 17:05:32  abrtd: DUP_OF_DIR:
> /var/spool/abrt/ccpp-2013-03-08-15:47:29-20599
> Mar  8 17:05:32  abrtd: Corrupted or bad directory
> '/var/spool/abrt/ccpp-2013-03-08-17:05:23-21781', deleting
> 
> The problem is that on my spacewalk server, I only have the first data
> uploaded, and the crash count is not updated.
> 
> Is this the expected behavior?

No, this is not expected.

Does abrt correctly update content of the 'count' file in the crash directory?
(since this is the number that spacewalk-abrt --update-count will report
to the server).

-MZ

> 2013/3/8 Pierre Casenove <pcasenove at gmail.com>
> 
> > Thanks a lot for the information. Having this functionnality is great!
> > 
> > Pierre
> > 
> > 
> > 2013/3/8 Milan Zazrivec <mzazrivec at redhat.com>
> > 
> >> > Thanks a lot, it works now... but I noticed 2 points:
> >> > 1) I've launched another test to validate the path. Here is the logs I
> >> 
> >> had:
> >> > Mar  8 15:46:44 abrtd: Directory 'ccpp-2013-03-08-15:46:44-20301'
> >> 
> >> creation
> >> 
> >> > detected
> >> > Mar  8 15:46:44  abrt[20327]: Saved core dump of pid 20301 (/bin/bash)
> >> 
> >> to
> >> 
> >> > /var/spool/abrt/ccpp-2013-03-08-15:46:44-20301 (413696 bytes)
> >> > Mar  8 15:46:56  abrtd: Sending an email...
> >> > Mar  8 15:46:56  abrtd: Email was sent to: root at localhost
> >> > Mar  8 15:46:56  abrtd: Duplicate: UUID
> >> > Mar  8 15:46:56 abrtd: DUP_OF_DIR:
> >> > /var/spool/abrt/ccpp-2013-03-07-15:29:42-3952
> >> > Mar  8 15:46:56 abrtd: Corrupted or bad directory
> >> > '/var/spool/abrt/ccpp-2013-03-08-15:46:44-20301', deleting
> >> 
> >> Yes, when abrtd detects a duplicate crash, it deletes its crash
> >> directory and
> >> increases the 'count' in the original directory (the one that this crash
> >> is
> >> a duplicate of).
> >> 
> >> > The folder "/var/spool/abrt/ccpp-2013-03-07-15:29:42-3952" is the one
> >> 
> >> that
> >> 
> >> > were created when spacewalk-abrt crashed.
> >> > When I've generated a new crash today with your patch, spacewalk-abrt
> >> > couldn't upload the data and deleted this folder.
> >> 
> >> abrtd won't execute spacewalk-abrt --report $DUMP_DIR when it detects
> >> duplicates (i.e. it won't re-upload the whole stuff for duplicates).
> >> Instead
> >> it launches spacewalk-abrt --update-count $DUMP_DIR which just increases
> >> the count number.
> >> 
> >> > I've deleted folder "/var/spool/abrt/ccpp-2013-03-07-15:29:42-3952"
> >> > and regenerated a crash... everything was working
> >> > 
> >> > 2) sosreport.tar.xz has not been uploaded to my spacewalk server. The
> >> 
> >> file
> >> 
> >> > size limit is set to 2048 and the file is 1.3MB.
> >> 
> >> This is correct. 2048 is a size limit in bytes (as stated on the org
> >> config
> >> web page).
> >> 
> >> -MZ
> >> 
> >> > 2013/3/8 Milan Zazrivec <mzazrivec at redhat.com>
> >> > 
> >> > > > Hello list,
> >> > > > I've upgraded my spacewalk server to version 1.9 (rhel 5 pgsql).
> >> > > > I've updated spacewalk client binairies on a test client, and
> >> 
> >> installed
> >> 
> >> > > > spacewalk-abrt package.
> >> > > > I've rebooted both the server and the client.
> >> > > > 
> >> > > > To test abrt crash upload functionnality, i've launched the
> >> 
> >> following
> >> 
> >> > > > compex script in background:
> >> > > > #!/bin/bash
> >> > > > 
> >> > > > while :
> >> > > > do
> >> > > > 
> >> > > >    sleep 1
> >> > > > 
> >> > > > done
> >> > > > 
> >> > > > Than, I killed it using kill -SIGSEGV 3952
> >> > > > 
> >> > > > As expected, abrt detected the crash... but then spacewalk
> >> > > > crashed! Mar  7 15:29:42  abrt[3967]: Saved core dump of pid 3952
> >> 
> >> (/bin/bash) to
> >> 
> >> > > > /var/spool/abrt/ccpp-2013-03-07-15:29:42-3952 (413696 bytes)
> >> > > > Mar  7 15:29:42  abrtd: Directory 'ccpp-2013-03-07-15:29:42-3952'
> >> > > 
> >> > > creation
> >> > > 
> >> > > > detected
> >> > > > Mar  7 15:29:50  abrtd: Sending an email...
> >> > > > Mar  7 15:29:50  abrtd: Email was sent to: root at localhost
> >> > > > Mar  7 15:29:51 abrtd: New problem directory
> >> > > > /var/spool/abrt/ccpp-2013-03-07-15:29:42-3952, processing
> >> > > > Mar  7 15:29:51  abrtd: An error has occurred:
> >> > > > Mar  7 15:29:51  abrtd: Error communicating with server. The
> >> > > > message
> >> 
> >> > > > was: Mar  7 15:29:51  abrtd: While running 'abrt.create_crash':
> >> caught
> >> 
> >> > > > Mar  7 15:29:51 tu-spa-d13 abrtd: exceptions.TypeError : expected
> >> > > > a character buffer object
> >> > > > Mar  7 15:29:51 tu-spa-d13 abrtd:
> >> > > > Mar  7 15:29:51 tu-spa-d13 abrtd: See /var/log/up2date for more
> >> > > 
> >> > > information
> >> > > 
> >> > > > In up2date log, I have:
> >> > > > <class 'up2date_client.up2dateErrors.CommunicationError'>: Error
> >> > > > communicating with server. The message was:
> >> > > > While running 'abrt.create_crash': caught
> >> > > > exceptions.TypeError : expected a character buffer object
> >> > > > 
> >> > > > 
> >> > > > I've checked httpd error log, I found this:
> >> > > > [Thu Mar 07 15:29:51 2013] [error] Exception Handler Information
> >> > > > [Thu Mar 07 15:29:51 2013] [error] Traceback (most recent call
> >> 
> >> last):
> >> > > > [Thu Mar 07 15:29:51 2013] [error]   File
> >> 
> >> "/usr/lib/python2.4/site-packages/spacewalk/server/apacheRequest.py",
> >> 
> >> > > line
> >> > > 
> >> > > > 122, in call_function
> >> > > > [Thu Mar 07 15:29:51 2013] [error]     response = apply(func,
> >> 
> >> params)
> >> 
> >> > > > [Thu Mar 07 15:29:51 2013] [error]   File
> >> > > > "/usr/share/rhn/server/handlers/xmlrpc/abrt.py", line 170, in
> >> > > 
> >> > > create_crash
> >> > > 
> >> > > > [Thu Mar 07 15:29:51 2013] [error]     server_crash_dir =
> >> > > > get_crash_path(str(server_org_id), str(self.server_id),
> >> > > > crash_data['crash']) [Thu Mar 07 15:29:51 2013] [error]   File
> >> > > > "/usr/lib/python2.4/site-packages/spacewalk/server/rhnLib.py",
> >> > > > line 218,
> >> > > 
> >> > > in
> >> > > 
> >> > > > get_crash_path
> >> > > > [Thu Mar 07 15:29:51 2013] [error]     if _is_secure_path(path):
> >> > > > [Thu Mar 07 15:29:51 2013] [error]   File
> >> > > > "/usr/lib/python2.4/site-packages/spacewalk/server/rhnLib.py",
> >> > > > line 211,
> >> > > 
> >> > > in
> >> > > 
> >> > > > _is_secure_path
> >> > > > [Thu Mar 07 15:29:51 2013] [error]     return not
> >> 
> >> path.startswith(('/',
> >> 
> >> > > > '../'))
> >> > > > [Thu Mar 07 15:29:51 2013] [error] TypeError: expected a character
> >> > > > buffer object
> >> > > > 
> >> > > > 
> >> > > > What am I doing wrong?
> >> > > 
> >> > > Heya -- you're not doing anything wrong. This is a valid bug in
> >> > > spacewalk-backend which will show on a RHEL-5 Spacewalk (server)
> >> > > only.
> >> 
> >> > > This is the fix you need on your server:
> >> http://git.fedorahosted.org/cgit/spacewalk.git/commit/?h=SPACEWALK-1.9&i
> >> d
> >> 
> >> > > =1d43a4da660df4ba4c8b4c85339bfbda65c0d049
> >> > > 
> >> > > These are spacewalk-backend packages for Spacewalk 1.9 containing
> >> > > the fix above:
> >> > > 
> >> > > http://koji.spacewalkproject.org/koji/buildinfo?buildID=30602
> >> > > 
> >> > > I'll try to get these packages to Spacewalk 1.9 repo in a near
> >> > > future.
> >> > > 
> >> > > Thank you for your report.
> >> > > -Milan Zázrivec
> >> > > 
> >> > > _______________________________________________
> >> > > Spacewalk-list mailing list
> >> > > Spacewalk-list at redhat.com
> >> > > https://www.redhat.com/mailman/listinfo/spacewalk-list
> >> 
> >> _______________________________________________
> >> Spacewalk-list mailing list
> >> Spacewalk-list at redhat.com
> >> https://www.redhat.com/mailman/listinfo/spacewalk-list

More information about the Spacewalk-list mailing list