Description of problem: When internal server error occurs during rhnpush, rhnpush printouts lot of meaningless xhtml directly to the console. Version-Release number of selected component (if applicable): Spacewalk nightly 2010-11-12 spacewalk-backend-1.2.69-1.fc13.noarch spacewalk-oracle-1.2.3-1.fc13.noarch How reproducible: always on ISE Steps to Reproduce: #Push packages to spacewalk. $ rhnpush -d $BUILDPATH/RPMS/noarch -c $TEST_CHANNEL \ -u $RHN_USER -p $RHN_PASS --nosig \ --server=http://$RHN_SERVER/APP Actual results: terminal is flooded with xhtml Expected results: As before, nice description of problem. Do not print xhtml to console. Additional info: Given xhtml is not valid.
Created attachment 459965 [details] How does the rhnpush output look
(In reply to comment #0) > Steps to Reproduce: > #Push packages to spacewalk. > $ rhnpush -d $BUILDPATH/RPMS/noarch -c $TEST_CHANNEL \ > -u $RHN_USER -p $RHN_PASS --nosig \ > --server=http://$RHN_SERVER/APP > > > Actual results: > terminal is flooded with xhtml > > Expected results: > As before, nice description of problem. Do not print xhtml to console. Does this mean that in the past, when the same ISE happened during rhnpush, you would get different error message? This ISE is presumably exactly what the server returns for this kind of requests, and we do not have any easy way to strip it off.
The format of the error message really does not matter.
(In reply to comment #4) > Does this mean that in the past, when the same ISE happened during rhnpush, you > would get different error message? Yes, the format was very different. Internal server error 500 Internal Server Error Error pushing package.rpm: Error Message: File-MD5sum Error Class Code: 500 Error Class Info: Missing HTTP header information (500) You can see record of original error message for ex. in bug 568291.
Note the "the same ISE happened" in my question. In that other case, the server was still in such a reasonable state that it could give you that nice "Error Message ... Error Class ... Info" message (body) back. In the ISE that you've hit now, you simply got the most generic ISE 500 page there was. Due to a bug in the server, it was not even able to generate that nicer (shorter, simpler) page. On the other hand, it gave you something which could nicely be used in browser. I propose to WONTFIX this, as the cause of the ISE was the problem, and that was address.
I see and agree with the WONTFIX. Thanks for explanation.