Description of problem: When trying to convert a patch cluster with solaris2mpm using rhnpush-5.0.0-7, the program outputs: Opening archive, this may take a while Writing patch-cluster-solaris-10_Recommended-20071019-1.sparc-solaris-patch-cluster.mpm and then just hangs with lots of cpu usage. It never completes (at least after 24 hours) but does generate 1 large mpm that doesn't seem to work properly. The same patch cluster converts properly with rhnpush-4.2.2-3. I've uploaded the patch cluster here: http://dually.rdu.redhat.com/pub/10_Recommended.zip Version-Release number of selected component (if applicable): rhnpush-5.0.0-7 How reproducible: Steps to Reproduce: 1. run 'solaris2mpm 10_Recommended.zip' using rhnpush-5.0.0-7 2. 3. Actual results: Patch cluster created is not usable and program never completes Expected results: program completes and usable patch cluster is generated.
What happens if you use the temporary directory argument? This appears to be that same old issue.
I'm not sure what you mean by same old issue? On my machine with 2.5 gigs free of space /tmp/ is on the same partition as the directory i was using, and even if i use solaris2mpm --tempdir=./temp/ 10_recommended.zip it still does not work. I've had a couple of complaints from support about this when working on solaris related things. (Granted it may have all been with the same patch cluster).
When solaris2mpm is called, it in turn calls _my_popen to run the zip command in a separate process, which ends up in an infinite while loop where we keep making the same call if there is a read descriptor.. Before fix: [prad ~]$ python /usr/share/rhn/rhnpush/solaris2mpm.py Desktop/10_Recommended.zip Opening archive, this may take a while select return values rd:[], wr:[], ex:[] Writing patch-cluster-solaris-10_Recommended-20071019-1.sparc-solaris-patch-cluster.mpm select return values rd:[<open file '<fdopen>', mode 'r' at 0x9370218>, <open file '<fdopen>', mode 'r' at 0x9370140>], wr:[], ex:[] select return values rd:[<open file '<fdopen>', mode 'r' at 0x9370218>, <open file '<fdopen>', mode 'r' at 0x9370140>], wr:[], ex:[] select return values rd:[<open file '<fdopen>', mode 'r' at 0x9370218>, <open file '<fdopen>', mode 'r' at 0x9370140>], wr:[], ex:[] select return values rd:[<open file '<fdopen>', mode 'r' at 0x9370218>, <open file '<fdopen>', mode 'r' at 0x9370140>], wr:[], ex:[] select return values rd:[<open file '<fdopen>', mode 'r' at 0x9370218>, <open file '<fdopen>', mode 'r' at 0x9370140>], wr:[], ex:[] ...and stays here for ever as the read descriptor is always the same.. After fix: [prad ~]$ python /usr/share/rhn/rhnpush/solaris2mpm.py Desktop/10_Recommended.zip Opening archive, this may take a while Writing patch-cluster-solaris-10_Recommended-20071019-1.sparc-solaris-patch-cluster.mpm Writing patch-solaris-118560-02-1.sparc-solaris-patch.mpm Writing patch-solaris-118712-18-1.sparc-solaris-patch.mpm Writing patch-solaris-118731-01-1.sparc-solaris-patch.mpm Writing patch-solaris-118815-06-1.sparc-solaris-patch.mpm ... [pkilambi@prad ~]$ and uploading the same mpm's works correctly: [prad ~]$ rhnpush --server=http://rlx-1-14.rhndev.redhat.com/APP --username=admin --password=dog8code -c solaris-136540 patch-solaris-125418-01-1.sparc-solaris-patch.mpm --force Connecting to http://rlx-1-14.rhndev.redhat.com/APP url is http://rlx-1-14.rhndev.redhat.com/PACKAGE-PUSH Result codes: 200 OK Computing md5sum and package Info .This may take sometime ... Uploading package patch-solaris-125418-01-1.sparc-solaris-patch.mpm Using POST request
Verified.
Solaris Management Build 5.1 Sat GA so Closed for Current Release.