TESTOUT.log is sent to server every 5 minutes - see rhts-test-runner.sh. Though it is growing, the file is send from offest 0 every time. High priority due to the data-overload is known to cause problems on LC/Scheduler. + the change will save us lot of bandwidth.
Testing the change.
Created attachment 443493 [details] rhts patch Following patch works fine in el6. Testing on 3 to 5.
Hi Marian, Slight issue when run on RHEL3. https://beaker-stage.app.eng.bos.redhat.com/logs/2010/29/1429/2690/26412///TESTOUT.log /distribution/install result: PASS metric: 441 Log: /tmp/tmp.ir1820 DMesg: /tmp/dmesg.log Traceback (most recent call last): File "/usr/bin/rhts-db-submit-result", line 450, in ? main() File "/usr/bin/rhts-db-submit-result", line 398, in main (result, dmesgerrors)=check_for_errors(dmesgfile) File "/usr/bin/rhts-db-submit-result", line 305, in check_for_errors dmesg_log_file = mktemp_file() File "/usr/bin/rhts-db-submit-result", line 58, in mktemp_file (fh, f) = tempfile.mkstemp(prefix='tmp.rhts-db-submit-result.', dir='/tmp') AttributeError: 'module' object has no attribute 'mkstemp' /bin/cat: /mnt/testarea/modprobe.log: No such file or directory I'm looking to see what can be done.
hmm assumed this was due to the new upload code. but I'm not sure now. Still investigating.
Created attachment 445840 [details] python 2.2 incompatibility patch There is not much to investigate: documentation [1] says: > New in version 2.3. RHEL3 uses Python2.2. [1] http://docs.python.org/library/tempfile.html#tempfile.mkstemp7 Proposed patch: let python<=2.2 use mktemp and all above mkstemp. IIUC mkstemp was introduced accidentally.
Forget about previous comment. I have to admit embarrassingly I was not right: - mkstemp has a purpose: mktemp does not create the file, just returns a new name. - and worse: my "fix" is not python 2.2. compatible: mktemp change was not documented. I will not rely on documentation anymore... Embracing Bill's change which I was not aware of. The fix for original issue works just fine: see J:1438, J:1439 and J:1440. I will retest the changes on el3 which is not verified yet.
After some aborted jobs it's verified on el3 as well. It's all on rhts.git's origin/master now.
Did you notice I already committed a patch for python2.2 to deal with mkstemp?
this has been pushed