Advisory: ========== This update of rpm fixes several issues - Fix not all %transfiletriggerpostun triggers executing (rhbz#1514085) - Fix file lists getting fed to file triggers multiple times (rpm#370) Last but not least, a no more needed patch was removed (an upstream fix already covered the issue since previous update -- it was for bug #18797)
Assignee: bugsquad => qa-bugs
Any suggestion on how to test this update? I did some testing with: - rpm -Va - rpm -qia but don't know how to test the actual fixes.
CC: (none) => mageia
I've seen this error on a few installs since installing the rpm update, though it doesn't seem to interfere with the actual installations ... 25/25: certbot-apache ############################################################################################################## perl: backend/dbiset.c:62: dbiIndexSetUniq: Assertion `set->count > 0' failed. Aborted (core dumped) It doesn't seem to be related to what's being installed.
Keywords: (none) => feedbackCC: (none) => davidwhodgins
I'm also seeing messages such as ... BDB2053 Freeing read locks for locker 0x871: 9841/140032927323904 Again, the install works ok despite the msg.
Also just noticed, the packages are not being deleted from /var/cache/urpmi/rpms after successful installation.
Re comment 2: same here for the perl and core dump message.
CC: (none) => tarazed25
Re comments 3 and 4: and the same here for those two observations.
Hi. When using mock, the command that reports the error is python3. Re, commen 2. Also seen the same as comment 3. I've also reported this to dev@. Cheers, Stig
CC: (none) => smelror
Ran a couple of experiments with rpm and local rpm files and saw no sign of the errors so the cause definitely lies outside rpm.
Re comment 8. Maybe disregard that because urpmi does not report the errors for local rpms either.
Extra tests on Mageia 6 with local RPMs. $ rpm -qilp oneplay-dvd-1.1.3-1.x86_64.rpm Name : oneplay-dvd Version : 1.1.3 Release : 1 Architecture: x86_64 Install Date: (not installed) Group : Applications/Internet Size : 26139454 License : Proprietary Signature : (none) Source RPM : oneplay-dvd-1.1.3-1.src.rpm Build Date : Fri 05 Jun 2015 12:27:51 BST Build Host : ubuntu1004-64.vmbuild.lan Relocations : /opt/oneplay-dvd Packager : Fluendo S.A. <support@fluendo.com> Vendor : Fluendo S.A. URL : http://www.fluendo.com/ Summary : ONEPLAY DVD player Description : Fluendo DVD Player is a software application specially designed to reproduce DVD on Linux/Unix platforms, which provides end users with high quality standards. * Full DVD Playback * DVD Menu support ..................................... $ sudo rpm -i mplayer-skins-1.8-1.nodist.rf.noarch.rpm seemed to go OK. mga6 tkimg package already installed so this was expected to fail. $ rpm -i --test tkimg-1.4-20.fc21.x86_64.rpm file /usr/lib64/tcl8.6/Img1.4/libjpegtcl8.2.so from install of tkimg-1.4-20.fc21.x86_64 conflicts with file from package tkimg-1.4-7.mga6.x86_64 file /usr/lib64/tcl8.6/Img1.4/libpngtcl1.4.3.so from install of tkimg-1.4-20.fc21.x86_64 conflicts with file from package tkimg-1.4-7.mga6.x86_64 .....................................
(In reply to Len Lawrence from comment #8) > Ran a couple of experiments with rpm and local rpm files and saw no sign of > the errors so the cause definitely lies outside rpm. Given that the packages install ok, the rpm db is updated (except for the locks being left in place), but the cache files are not deleted, my guess is that the problem is either in the freeing of the locks, or in the cleaning of the cache. When installing from local files, they do not get copied to, or deleted from the cache. As the errors don't show up with local rpm files, it's most likely the cleaning from the cache, causing the failure.
Comment 11: "When installing from local files, they do not get copied to, or deleted from the cache. As the errors don't show up with local rpm files, it's most likely the cleaning from the cache, causing the failure." Exactly my conclusion.
Ref comment 3. I just got this when I ran urpmi to update my mga6 system. In case it means anything to anyone. BDB2053 Freeing read locks for locker 0x8f: 28018/139963843876608 BDB2053 Freeing read locks for locker 0x91: 28018/139963843876608 BDB2053 Freeing read locks for locker 0x92: 28018/139963843876608 BDB2053 Freeing read locks for locker 0x93: 28018/139963843876608 BDB2053 Freeing read locks for locker 0x94: 28018/139963843876608 BDB2053 Freeing read locks for locker 0x95: 28018/139963843876608 BDB2053 Freeing read locks for locker 0x96: 28018/139963843876608 BDB2053 Freeing read locks for locker 0x97: 28018/139963843876608 BDB2053 Freeing read locks for locker 0x98: 28018/139963843876608 BDB2053 Freeing read locks for locker 0x99: 28018/139963843876608 BDB2053 Freeing read locks for locker 0x9a: 28018/139963843876608 BDB2053 Freeing read locks for locker 0x9b: 28018/139963843876608 BDB2053 Freeing read locks for locker 0x9c: 28018/139963843876608 BDB2053 Freeing read locks for locker 0x9d: 28018/139963843876608 BDB2053 Freeing read locks for locker 0x9e: 28018/139963843876608 BDB2053 Freeing read locks for locker 0x9f: 28018/139963843876608 BDB2053 Freeing read locks for locker 0xa0: 28018/139963843876608 BDB2053 Freeing read locks for locker 0xa1: 28018/139963843876608 BDB2053 Freeing read locks for locker 0xa2: 28018/139963843876608 Proceeded to install without the message from comment 2. Cheers, Stig
I've dropped this one from the mirrors as it's broken, and as a priority upgrade and screws with other QA tests
CC: (none) => tmb
CC: (none) => bequimao.de
To downgrade rpm ... urpmi --downgrade \ lib64rpm7-4.13.0.2-3.1.mga6 \ lib64rpmbuild7-4.13.0.2-3.1.mga6 \ lib64rpmsign7-4.13.0.2-3.1.mga6 \ python2-rpm-4.13.0.2-3.1.mga6 \ python3-rpm-4.13.0.2-3.1.mga6 \ rpm-4.13.0.2-3.1.mga6 \ rpm-build-4.13.0.2-3.1.mga6
Re-assigning back to Thierry.
Assignee: qa-bugs => thierry.vignaud
rpm has been further updated Advisory: ========== This update of rpm to 4.13.1 mainly fixes several issues regarding file triggers: - Fix file lists getting fed to file triggers multiple times (rpm#370) - Fix not all %transfiletriggerpostun file triggers executing (RhBz#1514085) - Fix file triggers executing before file fingerprinting - Fix file triggers firing on non-installed files - Fix file signatures failing on hardlinked files (rpm#333) See http://rpm.org/wiki/Releases/4.13.1 for the full details
Source RPM: rpm-4.13.0.2-3.2.mga6 => rpm-4.13.1-3.2.mga6Assignee: thierry.vignaud => qa-bugs
Keywords: feedback => (none)
Mageia 6 :: x86_64 Updated rpm packages: - lib64rpm7-4.13.1-3.2.mga6.x86_64 - lib64rpmbuild7-4.13.1-3.2.mga6.x86_64 - python2-rpm-4.13.1-3.2.mga6.x86_64 - python3-rpm-4.13.1-3.2.mga6.x86_64 - rpm-4.13.1-3.2.mga6.x86_64 - rpm-build-4.13.1-3.2.mga6.x86_64 Installed local rpm files. $ sudo rpm -i oneplay-dvd-1.1.3-1.x86_64.rpm error: Failed dependencies: CodeMeter is needed by oneplay-dvd-1.1.3-1.x86_64 $ sudo rpm -i CodeMeter-5.21.1487-502.x86_64.rpm oneplay-dvd-1.1.3-1.x86_64.rpm Starting CodeMeter Server: [ OK ] gtk-update-icon-cache: Cache file created successfully. Tested urpmi: # urpmi celestia To satisfy dependencies, the following packages are going to be installed: Package Version Release Arch (medium "Core Release") celestia 1.6.1 18.mga6 x86_64 lib64gtkglext-1.0_0 1.2.0 21.mga6 x86_64 lib64pangox1.0_0 0.0.2 6.mga6 x86_64 66MB of additional disk space will be used. 32MB of packages will be retrieved. Proceed with the installation of the 3 packages? (Y/n) $MIRRORLIST: media/core/release/lib64gtkglext-1.0_0-1.2.0-21.mga6.x86_64.rpm $MIRRORLIST: media/core/release/celestia-1.6.1-18.mga6.x86_64.rpm $MIRRORLIST: media/core/release/lib64pangox1.0_0-0.0.2-6.mga6.x86_64.rpm installing celestia-1.6.1-18.mga6.x86_64.rpm lib64pangox1.0_0-0.0.2-6.mga6.x86_64.rpm lib64gtkglext-1.0_0-1.2.0-21.mga6.x86_64.rpm from /var/cache/urpmi/rpms Preparing... ############################################# 1/3: lib64pangox1.0_0 ############################################# 2/3: lib64gtkglext-1.0_0 ############################################# 3/3: celestia ############################################# # exit $ sudo rpm -qilp oneplay-dvd-1.1.3-1.x86_64.rpm Name : oneplay-dvd Version : 1.1.3 Release : 1 Architecture: x86_64 Install Date: (not installed) Group : Applications/Internet Size : 26139454 License : Proprietary Signature : (none) Source RPM : oneplay-dvd-1.1.3-1.src.rpm ............. $ urpmq --not-available lib64torrent-rasterbar7-0.16.18-1.3.mga5.x86_64 lib64kf5jsapi5-5.42.0-1.mga6.x86_64 lib64camel1.2_52-3.13.90-1.mga5.x86_64 lib64xft-gir2.0-1.54.1-1.mga6.x86_64 lib64kf5windowsystem5-5.42.0-1.mga6.x86_64 ..................... This all looks fine.
Whiteboard: (none) => MGA6-64-OK
on mga6-32 plasma-5.12 4.14.30-desktop in a vbox VM packages updated cleanly: - librpm7-4.13.1-3.2.mga6.i586 - python2-rpm-4.13.1-3.2.mga6.i586 - python3-rpm-4.13.1-3.2.mga6.i586 - rpm-4.13.1-3.2.mga6.i586 $ urpmq -i gedit # urpmi gedit both responded correctly without errors, using my ftp mirror gedit and dependencies were installed and gedit can be launched and used: $ rpm -q gedit gedit-3.22.0-4.mga6 looks OK on mga6-32
CC: (none) => jimWhiteboard: MGA6-64-OK => MGA6-64-OK MGA6-32-OK
Advisory uploaded from comment 17 + bug RPMs list. I have seen some of the errors noted, like c13 and crashes at the end of rpm'ing (down to our own software, I suspect). @James: thanks. I was just about to ask for a 32-bit test too!
Keywords: (none) => advisory, validated_updateCC: (none) => sysadmin-bugs
(In reply to Lewis Smith from comment #20) > Advisory uploaded from comment 17 + bug RPMs list. > I have seen some of the errors noted, like c13 and crashes at the end of > rpm'ing (down to our own software, I suspect). Wait, what ? Do you see the errors mentioned in c13 with the 4.13.1-3.2.mga6 packages ?
Keywords: validated_update => (none)
With rpm-4.13.0.2-3.2.mga6 I have just seen the c13 error, followed by a crash, after urpmi-ing something. To try after the update.
CC: (none) => lewyssmith
One nice test is to install this update before trying the huge plasma/qt/... stack... if it survives that, it should hopefully be stable for all users
(In reply to Thomas Backlund from comment #21) > Do you see the errors mentioned in c13 with the 4.13.1-3.2.mga6 packages ? I do not think so. I must have pulled in the updated rpm packages as part of my mass update experiments: lib64rpm7-4.13.1-3.2.mga6 lib64rpmbuild7-4.13.1-3.2.mga6 lib64rpmsign7-4.13.1-3.2.mga6 python2-rpm-4.13.1-3.2.mga6 python3-rpm-4.13.1-3.2.mga6 rpm-4.13.1-3.2.mga6 Using this system now, a couple of urpmi operations did not show any error. (In reply to Thomas Backlund from comment #23) > One nice test is to install this update before trying the huge plasma/qt/... > stack... if it survives that, it should hopefully be stable for all users I have done just that, Thomas. See https://bugs.mageia.org/show_bug.cgi?id=22656#c23 https://bugs.mageia.org/show_bug.cgi?id=22656#c25 It is true that updated rpm was part of this global update, but (at least from MCC) put 'up front' to happen before the rest. If you are happy, please re-instate 'validated update'.
Ah, you did full updates with updates_testing enabled. Then yes, when done like that, rpm gets installed first as its a priority upgrade... and it happily chewed through 640+ packages... and gave you a working system... So it got a good test set of normal rpm installs/upgrades... Thanks for the extra info... However I dont see any rpm-build tests, so I'm going to push some local rpm buildtests here before validating it again
rpm building and signing still works, and so does iurt builds... so re-validating
Keywords: (none) => validated_update
An update for this issue has been pushed to the Mageia Updates repository. https://advisories.mageia.org/MGAA-2018-0058.html
Resolution: (none) => FIXEDStatus: NEW => RESOLVED