Join GitHub today
GitHub is home to over 28 million developers working together to host and review code, manage projects, and build software together.
Sign upVarious dom0 update errors on 2017-04-18 #2760
Comments
andrewdavidwong
added
bug
C: core
labels
Apr 18, 2017
andrewdavidwong
added this to the Release 3.2 updates milestone
Apr 18, 2017
This comment has been minimized.
Show comment
Hide comment
This comment has been minimized.
marmarek
Apr 18, 2017
Member
Traceback (most recent call last): File "/bin/qvm-ls", line 287, in main() File "/bin/qvm-ls", line 280, in main print s IOError: [Errno 32] Broken pipe
This one is probably safe to ignore, comes from qvm-ls -k|grep -q ... and grep -q terminates on first match.
This kernel version is used by at least one VM, cannot remove error:
%preun(kernel-qubes-vm-1000:4.4.14-11.pvops.qubes.x86_64) scriptlet failed, exit status 1
Error in PREUN scriptlet in rpm package kernel-qubes-vm
Error in PREUN scriptlet in rpm package kernel-qubes-vm
Erasing : kernel-qubes-vm-1000:4.4.14-11.pvops.qubes.x86_64
This is the real problem - even when PREUN script fails, rpm proceed to remove the package. Apparently this is fixed upstream (rpm-software-management/rpm@3eb469b), and the fix is part of rpm-4.13.0. But this package stuck in Fedora updates-testing (https://bodhi.fedoraproject.org/updates/FEDORA-2016-0906f64ec8) and fc23 hit EOL then. So we're left with broken rpm-4.13.0-rc1 :/
BTW it's interesting that "release candidate" version of critical system component was allowed, but actual stable release was hold.
This one is probably safe to ignore, comes from
This is the real problem - even when PREUN script fails, rpm proceed to remove the package. Apparently this is fixed upstream (rpm-software-management/rpm@3eb469b), and the fix is part of rpm-4.13.0. But this package stuck in Fedora updates-testing (https://bodhi.fedoraproject.org/updates/FEDORA-2016-0906f64ec8) and fc23 hit EOL then. So we're left with broken rpm-4.13.0-rc1 :/ |
This comment has been minimized.
Show comment
Hide comment
This comment has been minimized.
andrewdavidwong
Apr 18, 2017
Member
What about:
sed: can't read /etc/sysconfig/prelink: No such file or directory
Redirecting to /bin/systemctl start xenstored.service
Any problem there?
|
What about:
Any problem there? |
This comment has been minimized.
Show comment
Hide comment
This comment has been minimized.
|
No, safe to ignore. I'll mute those messages. |
This comment has been minimized.
Show comment
Hide comment
This comment has been minimized.
andrewdavidwong
Apr 19, 2017
Member
No, safe to ignore. I'll mute those messages.
Any idea how the other errors (including #2757) managed to land in stable? Do we still not have enough people testing dom0 updates?
Any idea how the other errors (including #2757) managed to land in stable? Do we still not have enough people testing dom0 updates? |
This comment has been minimized.
Show comment
Hide comment
This comment has been minimized.
marmarek
Apr 19, 2017
Member
Any idea how the other errors (including #2757) managed to land in stable?
That problem apply only to systems where kernel version was manually set to some explicit value. By default it's "default" and follow newest installed version. That package was in testing for almost a month (QubesOS/updates-status#16).
Do we still not have enough people testing dom0 updates?
Apparently... Or maybe everyone testing updates had kernel set to "default"?
BTW what was the reason to set kernel version manually to some specific value in the first place? Some other broken update in the past?
That problem apply only to systems where kernel version was manually set to some explicit value. By default it's "default" and follow newest installed version. That package was in testing for almost a month (QubesOS/updates-status#16).
Apparently... Or maybe everyone testing updates had kernel set to "default"? |
This comment has been minimized.
Show comment
Hide comment
This comment has been minimized.
andrewdavidwong
Apr 19, 2017
Member
BTW what was the reason to set kernel version manually to some specific value in the first place? Some other broken update in the past?
Could be. I don't recall ever intentionally changing it.
Could be. I don't recall ever intentionally changing it. |
This comment has been minimized.
Show comment
Hide comment
This comment has been minimized.
mfc
Apr 19, 2017
Member
BTW what was the reason to set kernel version manually to some specific value in the first place? Some other broken update in the past?
Could be. I don't recall ever intentionally changing it.
yeah I've never changed them myself.
yeah I've never changed them myself. |
This comment has been minimized.
Show comment
Hide comment
This comment has been minimized.
mvermaes
Apr 19, 2017
In case it helps - I have never changed kernel from default either, but am using a custom template based on fedora-24-minimal. All AppVMs based on this fedora-24 template were not set to kernel 4.4.55-11 after rebooting for this latest dom0 update.
Using the procedure in #2757 to change them to the new default kernel corrected the issue.
The AppVMs I have that are based on the default fedora-23 template were using the new kernel already, these didn't need to be reconfigured.
mvermaes
commented
Apr 19, 2017
|
In case it helps - I have never changed kernel from default either, but am using a custom template based on fedora-24-minimal. All AppVMs based on this fedora-24 template were not set to kernel 4.4.55-11 after rebooting for this latest dom0 update. Using the procedure in #2757 to change them to the new default kernel corrected the issue. The AppVMs I have that are based on the default fedora-23 template were using the new kernel already, these didn't need to be reconfigured. |
This comment has been minimized.
Show comment
Hide comment
This comment has been minimized.
0spinboson
Apr 21, 2017
Marek: "Apparently... Or maybe everyone testing updates had kernel set to "default"?"
In my case, I never noticed it before because I still had 4.4.11 installed alongside 4.8.12, so I never got any of the 4.4 updates (because they're ignored), and I was only confronted with the problem when I manually removed 4.4.11 yesterday.
0spinboson
commented
Apr 21, 2017
•
|
Marek: "Apparently... Or maybe everyone testing updates had kernel set to "default"?" In my case, I never noticed it before because I still had 4.4.11 installed alongside 4.8.12, so I never got any of the 4.4 updates (because they're ignored), and I was only confronted with the problem when I manually removed 4.4.11 yesterday. |
andrewdavidwong commentedApr 18, 2017
Qubes OS version (e.g.,
R3.2):R3.2@marmarek: AFAIK, we don't really have a place for this sort of thing, so I'm just reporting the output here:
Related issues:
Kernel bug already reported in #2757.