Not sure which list is most appropriate, but starting with test to see if this is something which would be expected to be caught while a package was in updates-testing:
I just ran dnf upgrade on two different F25 systems, and on both dnf now crashes with a segmentation fault if run sudo or under su. Running as normal user works OK, but of course from my normal user account I can't actually update packages. RPM also crashes, so I don't know if both are crashing independently, but I suspect that the update to libdb 5.3.28-21 is the culprit (the other package updated at the same time was google-noto-emoji-fonts, seems unlikely to be related).
Am I completely dead in the water for now? After the update tracer reported that nothing needed to be restarted, but running tracer now has this error:
# tracer error: rpmdb: BDB0113 Thread/process 14829/140660169317184 failed: BDB1507 Thread died in Berkeley DB library error: db5 error(-30973) from dbenv->failchk: BDB0087 DB_RUNRECOVERY: Fatal error, run database recovery error: cannot open Packages index using db5 - (-30973) error: cannot open Packages database in /var/lib/rpm Traceback (most recent call last): File "/usr/bin/tracer", line 34, in <module> tracer.main.run() File "/usr/lib/python3.5/site-packages/tracer/main.py", line 45, in run return router.dispatch() File "/usr/lib/python3.5/site-packages/tracer/resources/router.py", line 52, in dispatch controller = DefaultController(self.args, self.packages) File "/usr/lib/python3.5/site-packages/tracer/controllers/default.py", line 62, in __init__ self.applications = self.tracer.trace_affected(self._user(args.user)) File "/usr/lib/python3.5/site-packages/tracer/resources/tracer.py", line 96, in trace_affected for file in self._PACKAGE_MANAGER.package_files(package.name): File "/usr/lib/python3.5/site-packages/tracer/resources/PackageManager.py", line 55, in package_files return self.package_managers[0].package_files(pkg_name) File "/usr/lib/python3.5/site-packages/tracer/packageManagers/dnf.py", line 34, in package_files if self._is_installed(pkg_name): File "/usr/lib/python3.5/site-packages/tracer/packageManagers/rpm.py", line 151, in _is_installed mi = ts.dbMatch('name', pkg_name) _rpm.error: rpmdb open failed
Open to debugging suggestions. Normally I would run dnf downgrade, but that obviously doesn't work when rpm is crashing.
Oddly enough, rpm when run as a normal user works OK, at least enough to report which files are in the libdb package: $ rpm -ql libdb /usr/lib64/libdb-5.3.so /usr/lib64/libdb-5.so /usr/share/doc/libdb /usr/share/doc/libdb/README /usr/share/licenses/libdb /usr/share/licenses/libdb/LICENSE /usr/share/licenses/libdb/lgpl-2.1.txt /usr/lib/libdb-5.3.so /usr/lib/libdb-5.so /usr/share/doc/libdb /usr/share/doc/libdb/README /usr/share/licenses/libdb /usr/share/licenses/libdb/LICENSE /usr/share/licenses/libdb/lgpl-2.1.txt
When run sudo rpm crashes:
]$ sudo rpm -ql libdb error: rpmdb: BDB0113 Thread/process 14829/140660169317184 failed: BDB1507 Thread died in Berkeley DB library error: db5 error(-30973) from dbenv->failchk: BDB0087 DB_RUNRECOVERY: Fatal error, run database recovery error: cannot open Packages index using db5 - (-30973) error: cannot open Packages database in /var/lib/rpm error: rpmdb: BDB0113 Thread/process 14829/140660169317184 failed: BDB1507 Thread died in Berkeley DB library error: db5 error(-30973) from dbenv->failchk: BDB0087 DB_RUNRECOVERY: Fatal error, run database recovery error: cannot open Packages index using db5 - (-30973) error: cannot open Packages database in /var/lib/rpm package libdb is not installed
I saw the message from Eric Blake and did confirm that sudo rpm --rebuilddb will recover. I did not have to delete the corrupted database file first, rpm rebuilt the db even without that step.
On Fri, 2017-06-09 at 16:14 -0500, Chris Caudle wrote:
Not sure which list is most appropriate, but starting with test to see if this is something which would be expected to be caught while a package was in updates-testing:
This update was kind of a special case, as it was required to fix a blocker bug in F26 (under certain circumstances, an update to F24 / F25 can be required to clear a 'blocker' condition for F26). The update has been undergoing testing by folks on the relevant bug reports all week, but it seems like it never actually made it to updates-testing even though someone tried to push it early last week (possibly a Bodhi issue), so it more or less went straight to stable without really hitting u-t. Unfortunately it does seem to still have some issues with certain package updates causing RPM db issues :/
Fortunately rpm --rebuilddb does seem to clear it, so it's recoverable, but this is definitely turning into a real mess. Sorry about that. I've sent a -22 to u-t which reverts the offending changes (again), and posted a PSA to my blog and G+ about the potential DB corruption; if folks could help spread that around it'd be great.
Relevant bugs are:
https://bugzilla.redhat.com/show_bug.cgi?id=1394862 https://bugzilla.redhat.com/show_bug.cgi?id=1397087 https://bugzilla.redhat.com/show_bug.cgi?id=1460303