Description of problem: up2date failed while running rpm transaction during upgrade from RHEL4-U6/AS to RHEL4.7 through RHN.webqa Version-Release number of selected component (if applicable): up2date-4.7.1-1.el4 (since it's updated first and then restarted) How reproducible: Steps to Reproduce: 1. Install RHEL4-U6::AS, @everything, all languages 2. Configure and register to rhn.webqa.redhat.com 3. Subscribe to BETA channels 4. Import necessary GPG keys 5. run up2date -fu Actual results: After 400 (and something) packages are installed the update just stops and spits out a message: There was a fatal RPM install error. The message was: Failed running rpm transaction Expected results: update works without problems Additional info: last lines from /var/log/up2date are [Mon May 19 15:04:24 2008] up2date installing packages: [] [Mon May 19 15:04:25 2008] up2date Failed running rpm transaction - %pre %pro failure ?. running up2date manually gives: # up2date -fu Fetching all package list for channel: rhel-ia64-as-4... ######################################## Fetching all package list for channel: rhel-ia64-as-4-beta... ######################################## Fetching Obsoletes list for channel: rhel-ia64-as-4... Fetching Obsoletes list for channel: rhel-ia64-as-4-beta... Name Version Rel ---------------------------------------------------------- mstflint 1.3 1.el4 ia64 perftest 1.2 11.el4 ia64 tvflash 0.9.0 2.el4 ia64 Testing package set / solving RPM inter-dependencies... ######################################## There was a fatal RPM install error. The message was: Failed running rpm transaction
Created attachment 305958 [details] /var/log/up2date
Created attachment 305959 [details] console output captured with `script` - a bit messy
Pradeep, is it possible that we have 2 bugs here 1) Something is wrong with the packages (this bug) 2) RPM is not telling us what is wrong (see bug #437813 - rhel5 for similar situation) Thanks!
Hi - I suspect and expect that this issue is with wireshark, and not up2date. Most likely an rpm package issue which is preventing the rpm transaction from completing. up2date is just being nice to return an error that rpm hit. from 2nd attachment: 448:wireshark-gnome # ( 2%)## ( 4%)### ( 7%)#### ( 9%)##### ( 12%)###### ( 14%)####### ( 16%)######## ( 19%)######### ( 21%)########## ( 24%)########### ( 26%)############ ( 29%)############# ( 31%)############## ( 33%)################ ( 36%)################# ( 38%)################## ( 41%)################### ( 43%)#################### ( 46%)##################### ( 48%)###################### ( 50%)####################### ( 53%)######################## ( 55%)######################### ( 58%)########################## ( 60%)########################### ( 63%)############################ ( 65%)############################# Up2date able to install 447 packages, but not the 448 - 9.9 times out of 10 it is the package in question, and not up2date. I am willing to look at the issue with QA and devel for wireshark-gnome. I am on #satellite as cliff* or cperry*
Not wireshark-gnome - it seems, so far not able to determine what on this system used by QA is causing the issue though. Still investigating to figure out. Cliff
This is not wireshark. Tested the same system, same conditions but: 1) up2date wireshark 2) up2date -fu Same failure this time after installing xemacs-info.
From the log file: up2date installing packages: [] up2date Failed running rpm transaction - %pre %pro fa Any idea why there are no packages selected? (i.e. an empty transaction) Is that possible at all? Thanks!
John MAtthews was looking at this yesterday, ping jmatthews on #satellite or #satellite-devel for the insight/review he was doing
Filed as bug #447707: Circular obsoletes with the new renamed openib-* packages If there're not going to be any changes in up2date code base can we close this bug? Please advise.
Changing to blocker-rc because of comment #16.
Closing as NOTABUG. Exact issue reported as bug #447707 which will be used for future tracking.