Bug is already filed against the correct component. This is just an FYI.
On Wed, Jul 19, 2017 at 4:28 AM, Dusty Mabe dusty@dustymabe.com wrote:
Bug is already filed against the correct component. This is just an FYI.
You've got the wrong end of the stick, nothing provides libxenctrl.so.4.8 because xen was moved to 4.9 and hence qemu, and probably libvirt and others, needs to be rebuilt against it.
P
On 07/19/2017 04:12 AM, Peter Robinson wrote:
On Wed, Jul 19, 2017 at 4:28 AM, Dusty Mabe dusty@dustymabe.com wrote:
Bug is already filed against the correct component. This is just an FYI.
You've got the wrong end of the stick, nothing provides libxenctrl.so.4.8 because xen was moved to 4.9 and hence qemu, and probably libvirt and others, needs to be rebuilt against it.
So is it not a bug against qemu to do the update and rebuild?
Dusty
On 07/19/2017 04:12 AM, Peter Robinson wrote:
On Wed, Jul 19, 2017 at 4:28 AM, Dusty Mabe dusty@dustymabe.com wrote:
Bug is already filed against the correct component. This is just an FYI.
You've got the wrong end of the stick, nothing provides libxenctrl.so.4.8 because xen was moved to 4.9 and hence qemu, and probably libvirt and others, needs to be rebuilt against it.
So is it not a bug against qemu to do the update and rebuild?
You worded it differently to that, it was a known problem, they were having issues rebuilding due to the breakage in s390x, so it was already in progress.
On 07/19/2017 07:55 AM, Peter Robinson wrote:
You worded it differently to that, it was a known problem, they were having issues rebuilding due to the breakage in s390x, so it was already in progress.
Can you help explain to me what I did wrong?
AFAIK I found a problem and opened a bug against the correct component. I had no prior knowledge of ongoing efforts mainly because they weren't announced anywhere (at least I can't find it anywhere) and I looked at previous bugs against qemu before filing this one. Maybe one exists, but I wasn't looking for the right text?
Dusty
You worded it differently to that, it was a known problem, they were having issues rebuilding due to the breakage in s390x, so it was already in progress.
Can you help explain to me what I did wrong?
AFAIK I found a problem and opened a bug against the correct component. I had no prior knowledge of ongoing efforts mainly because they weren't announced anywhere (at least I can't find it anywhere) and I looked at previous bugs against qemu before filing this one. Maybe one exists, but I wasn't looking for the right text?
"dnf repoquery --whatprovides libxenctrl.so.4.8" told me the problem package was xen, looking in koji [1] showed that it was upgraded a few days ago, then the combination of qemu build failures since [2] (see build -4 and associated changelog) coupled with the fact we'd not had a compose in a few days provided the remainder of the info.
[1] https://koji.fedoraproject.org/koji/packageinfo?packageID=7 [2] https://koji.fedoraproject.org/koji/packageinfo?packageID=3685
On 07/19/2017 08:49 AM, Peter Robinson wrote:
You worded it differently to that, it was a known problem, they were having issues rebuilding due to the breakage in s390x, so it was already in progress.
Can you help explain to me what I did wrong?
AFAIK I found a problem and opened a bug against the correct component. I had no prior knowledge of ongoing efforts mainly because they weren't announced anywhere (at least I can't find it anywhere) and I looked at previous bugs against qemu before filing this one. Maybe one exists, but I wasn't looking for the right text?
"dnf repoquery --whatprovides libxenctrl.so.4.8" told me the problem package was xen, looking in koji [1] showed that it was upgraded a few days ago, then the combination of qemu build failures since [2] (see build -4 and associated changelog) coupled with the fact we'd not had a compose in a few days provided the remainder of the info.
[1] https://koji.fedoraproject.org/koji/packageinfo?packageID=7 [2] https://koji.fedoraproject.org/koji/packageinfo?packageID=3685
I agree maybe I could have done more, but I maintain that I didn't do anything incorrect. It would be great if we actually notified when there are issues with rawhide (or any compose really) so many different people don't have to investigate it.
Example. Oh crap compose X didn't build last night. I wonder what the problem is. Go to canonical location where discussion about compose X should happen and see if anyone added any notes there. Yes, I see dep issues around new version of xen and it's being worked on - great. Don't have to ask anyone or email anywhere.
In Bodhi there is one place to go to look at discussion for a particular rpm (update) and if it is working or not. It would be great if we had something like this for pungi.
Dusty
On Wed, Jul 19, 2017 at 2:02 PM, Dusty Mabe dusty@dustymabe.com wrote:
On 07/19/2017 08:49 AM, Peter Robinson wrote:
You worded it differently to that, it was a known problem, they were having issues rebuilding due to the breakage in s390x, so it was already in progress.
Can you help explain to me what I did wrong?
AFAIK I found a problem and opened a bug against the correct component. I had no prior knowledge of ongoing efforts mainly because they weren't announced anywhere (at least I can't find it anywhere) and I looked at previous bugs against qemu before filing this one. Maybe one exists, but I wasn't looking for the right text?
"dnf repoquery --whatprovides libxenctrl.so.4.8" told me the problem package was xen, looking in koji [1] showed that it was upgraded a few days ago, then the combination of qemu build failures since [2] (see build -4 and associated changelog) coupled with the fact we'd not had a compose in a few days provided the remainder of the info.
[1] https://koji.fedoraproject.org/koji/packageinfo?packageID=7 [2] https://koji.fedoraproject.org/koji/packageinfo?packageID=3685
I agree maybe I could have done more, but I maintain that I didn't do anything incorrect. It would be great if we actually notified when there are issues with rawhide (or any compose really) so many different people don't have to investigate it.
You mean like the rawhide report? It listed the problems: https://lists.fedoraproject.org/archives/list/devel@lists.fedoraproject.org/...
Example. Oh crap compose X didn't build last night. I wonder what the problem is. Go to canonical location where discussion about compose X should happen and see if anyone added any notes there. Yes, I see dep issues around new version of xen and it's being worked on - great. Don't have to ask anyone or email anywhere.
In Bodhi there is one place to go to look at discussion for a particular rpm (update) and if it is working or not. It would be great if we had something like this for pungi.
Yea, and rawhide doesn't have bodhi so reading the rawhide daily report is a good spot to see what you'd expect to be broken that day and a list of what changed and hence what might have caused the issue. It's probably in your inbox most mornings by the time you sit down.... if it's not there's probably a problem....
On 07/19/2017 09:10 AM, Peter Robinson wrote:
On Wed, Jul 19, 2017 at 2:02 PM, Dusty Mabe dusty@dustymabe.com wrote:
I agree maybe I could have done more, but I maintain that I didn't do anything incorrect. It would be great if we actually notified when there are issues with rawhide (or any compose really) so many different people don't have to investigate it.
You mean like the rawhide report? It listed the problems: https://lists.fedoraproject.org/archives/list/devel@lists.fedoraproject.org/...
I guess. Though it doesn't seem like much discussion happens there.
Example. Oh crap compose X didn't build last night. I wonder what the problem is. Go to canonical location where discussion about compose X should happen and see if anyone added any notes there. Yes, I see dep issues around new version of xen and it's being worked on - great. Don't have to ask anyone or email anywhere.
In Bodhi there is one place to go to look at discussion for a particular rpm (update) and if it is working or not. It would be great if we had something like this for pungi.
Yea, and rawhide doesn't have bodhi so reading the rawhide daily report is a good spot to see what you'd expect to be broken that day and a list of what changed and hence what might have caused the issue. It's probably in your inbox most mornings by the time you sit down.... if it's not there's probably a problem....
I'm fine with that being the place where we have a discussion if it's what we want. We do have to change culture and get people to have the discussion there, though. When something is broken and being worked on let's talk about it, link to bugs, etc.
Is this the wrong approach?
On 07/18/2017 11:28 PM, Dusty Mabe wrote:
Bug is already filed against the correct component. This is just an FYI.
This is resolved with last nights ostree compose:
https://koji.fedoraproject.org/koji/taskinfo?taskID=20609382
-
Dusty Mabe -
Peter Robinson