Description of problem:
I've tried to upgrade 15 of our workstations from FC3 to FC4. The nodes are
using a central yum cache via NFS. While the upgrading some of them had to be
aborted. They give the error msg on the bottom. Some of them has finished the
upgrade "successfully" but they do the same. Some of them (3) is ok.
I've tried to delete __db.* from /var/lib/rpm , but it wasn't help.
Version-Release number of selected component (if applicable):
Steps to Reproduce:
1. Upgrade from FC3 latest to FC4
2. Abort the upgrade
Repository extras already added, not adding again
Setting up Upgrade Process
Setting up repositories
livna 100% |=========================| 951 B 00:00
updates-released 100% |=========================| 951 B 00:00
base 100% |=========================| 1.1 kB 00:00
extras 100% |=========================| 1.1 kB 00:00
Reading repository metadata in from local files
Traceback (most recent call last):
File "/usr/bin/yum", line 15, in ?
File "/usr/share/yum-cli/yummain.py", line 72, in main
result, resultmsgs = do()
File "/usr/share/yum-cli/cli.py", line 428, in doCommands
File "/usr/share/yum-cli/cli.py", line 877, in updatePkgs
File "/usr/share/yum-cli/cli.py", line 78, in doRepoSetup
File "__init__.py", line 241, in doSackSetup
File "repos.py", line 283, in populateSack
File "sqlitecache.py", line 96, in getPrimary
File "sqlitecache.py", line 83, in _getbase
File "sqlitecache.py", line 79, in getDatabase
File "sqlitecache.py", line 239, in makeSqliteCacheFile
File "sqlitecache.py", line 150, in createTablesPrimary
File "sqlitecache.py", line 188, in createDbInfo
line 244, in execute
_sqlite.OperationalError: database is locked
An upgrade-ed system.
Additional info: rpm, yum are fresh, installed from the central repository.
the yum cache on nfs is where the problem is. You've got multiple clients
accessing the same sqlite db via nfs and at least one of them is locking the db
for a while.
That's the problem.
The problem is permanent, it's still exists since 2 weeks. Only one client
access the cache at the same time. If a client were upgraded successfully, they
can be again without problem.
right - b/c nfs locking of sqlite dbs is pretty much not going to work.
The resolution to this bug, most likely, is "don't do that"
Seth drove me to a workaround:
1) copy the whole NFS repo to the disk of the machine which needs upgrading
2) then yum upgrade
It's not too admin friendly, but works.
or just setup a local repo in nfs and use it from there.
either way - closing - wontfix