Description of problem:
When trying to install something, RPM/yum hang infinitely.
Version-Release number of selected component (if applicable):
# rpm -q rpm
# rpm -q yum
Always for some packages
Steps to Reproduce:
1. yum install xchat-gnome
2. yum will download the RPMs and try to install it
yum will hang on running the transaction (before installing the package).
yum should install the package.
I thought at first it was yum, but then I tried this:
# yumdoanloader xchat-gnome
xchat-gnome-0.18-12.fc9.x86_64.rpm | 1.6 MB 00:32
# rpm -ivh xchat-gnome-0.18-12.fc9.x86_64.rpm
attention: xchat-gnome-0.18-12.fc9.x86_64.rpm: Entête V3 DSA signature: NOKEY,
key ID 4f2a6fd2
And then rpm hangs...
Could it be that RPM can't install unsigned packages?
It doesn't happen on every packages (I was able to do an update with PackageKit,
ie with yum).
While it's hung, I get the following:
# top -b -n 1 -p 14008
top - 23:12:18 up 1:21, 4 users, load average: 1.03, 0.81, 0.87
Tasks: 1 total, 0 running, 1 sleeping, 0 stopped, 0 zombie
Cpu(s): 6.9%us, 1.7%sy, 0.0%ni, 89.0%id, 2.3%wa, 0.0%hi, 0.0%si, 0.0%st
Mem: 2052776k total, 1908300k used, 144476k free, 74004k buffers
Swap: 1051700k total, 0k used, 1051700k free, 1090340k cached
PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND
14008 root 20 0 344m 74m 8764 D 0.0 3.7 0:05.87 yum
(the %CPU and %MEM remains absolutely constant over time)
Ok, it's been happening since I installed Fedora 9 preview this week end.
I just switched my laptop on, both yum and rpm install the exact same packages
that were hanging previously :-/
That's really weird, I've been trying to "yum clean all", rebuild rpm db,
reboot, etc... And the issue was always the same. And now, it seems to be simply
Sorry for having opened the bug report.
This is likely a stale futex from the kernel.
Only a reboot can clear stale futexes unless robust mutexes are used.
There's a patch to use robust mutexes in rpm-4.4.9+ sources.
Changing version to '9' as part of upcoming Fedora 9 GA.
More information and reason for this action is here:
I'm wondering if this bug isn't somehow related to the fsync bug in Firefox,
which is related to the sqlite behaviour.
As yum uses sqlite IIRC, and as the symptoms are identical (the application
freezes, no CPU used, no / few memory used), I thought it might by a interesting
track to investigate.
However, as I said previously, this doesn't happen anymore...
Closing bug as the reporter has indicated that the issue is resolved.