Red Hat Bugzilla – Bug 503584
Cobbler's Apache config files overwrite PythonPath supplied by Satellite's Apache config files
Last modified: 2009-09-10 16:37:37 EDT
Description of problem: Since the earliest builds of 530, both developers and QA personnel have seen what appear to be deterministic 500 ISE errors upon accessing the XMLRPC API functionality of the Satellite. jsherrill and pkilambi believe that these issues derive from an issue wherein Cobbler's Apache configuration files override the PythonPath variable supplied by the Satellite's Apache config files. Regardless of whether this problem is the root cause of the deterministic ISEs, this configuration problem should be rectified. Version-Release number of selected component (if applicable): 530, 5/29 build How reproducible: Deterministic Steps to Reproduce: 1. View /etc/httpd/conf.d/cobbler.conf 2. Run a number of tests that hit Satellite's frontend and backend XMLRPC APIs (for instance, respectively: run a few api calls in a loop, register a lot of systems) Actual results: PythonPath overridden by cobbler's configuration file, 500 ISEs appear deterministically Expected results: PythonPath not overridden by cobbler's configuration file, 500 ISEs don't appear Additional info:
Prad - this needs reviewing. See if we can replicate and maybe fix. Cliff
So prad and I have discussed this at great length and while we think that it may have something to do with it, we can't remove those PythonPath lines from the cobbler config. Those were added because we were setting the PythonPath explicitly in our configs and when it got to the cobbler sections, apache (in a Directory config) wouldn't automatically include that directory in the PythonPath (which it would if we hadn't of set the PythonPath). So if we remove those lines cobbler will stop working. Those lines shouldn't automatically cause an issue because they are simply appending a directory to the python path and not overriding it. More investigation needs to be done.
will leave it on my plate for now..
*** Bug 504674 has been marked as a duplicate of this bug. ***
How i was able to reproduce this issue: 1. run "rhn-satellite restart" 2. Ensure that nothing else is trying to contact the satellite (no other clients or anything) 2. kickstart a box and wait for it to finish 3. AS SOON as it has rebooted, simply run "rhn_check" on that specific client about 10 times. 4. Should see an ISE within 10 tries I kickstarted using the cobbler boot iso but it shouldn't make a difference.
*** Bug 483156 has been marked as a duplicate of this bug. ***
VERIFIED on 6/19 build (as far as my testing can tell).
verified in stage on xen5.englab.brq
An advisory has been issued which should help the problem described in this bug report. This report is therefore being closed with a resolution of ERRATA. For more information on therefore solution and/or where to find the updated files, please follow the link below. You may reopen this bug report if the solution does not work for you. http://rhn.redhat.com/errata/RHEA-2009-1434.html