virt-manager fail to start on huawei arm server
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
libvirt (Ubuntu) |
Fix Released
|
Undecided
|
Unassigned | ||
Bionic |
Fix Released
|
Undecided
|
Unassigned | ||
Cosmic |
Fix Released
|
Undecided
|
Unassigned |
Bug Description
[Impact]
* Some arm systems can have unexpected high core_id's which breaks
libvirts arbitrary assumption to limit it at 4k
* Upstream fix drops this limit and makes use of libvirts ability to size
the bitsets automatically as needed
[Test Case]
* This is the hard part, as it needs special HW to be affected, but also
to verify the fix.
* We can ask the reporter to verify on his platform, but not 100% rely on
that in case he is unavailable.
[Regression Potential]
* The change is removing a limit which only hit very special HW. For this
special HW it is a fix, for all others it should be no functional
change.
* The one regression I could think of is that this dynamic bitmask
handling would have (not yet identified issues) that would affect other
cases (the feature is almost three years old now since lbivirt 1.3.3).
[Other Info]
* n/a
---
We met an issue when run virt-manager on huawei arm server with 18.04.1.
Error polling connection 'qemu:///system': internal error: Socket 5418 can't be handled (max socket is 4095)
Related branches
- Andreas Hasenack: Approve
- Canonical Server: Pending requested
- git-ubuntu developers: Pending requested
-
Diff: 426 lines (+380/-0)7 files modifieddebian/changelog (+9/-0)
debian/patches/series (+5/-0)
debian/patches/ubuntu/lp-1771662-1-util-fixing-wrong-assumption-that-PF-has-to-have-net.patch (+62/-0)
debian/patches/ubuntu/lp-1771662-2-util-Code-simplification.patch (+104/-0)
debian/patches/ubuntu/lp-1771662-3-util-Fix-for-NULL-dereference.patch (+74/-0)
debian/patches/ubuntu/lp-1771662-4-util-Fixing-invalid-error-checking-from-virPCIGetNet.patch (+35/-0)
debian/patches/ubuntu/lp-1811198-utils-Remove-arbitrary-limit-on-socket_id-core_id.patch (+91/-0)
- Andreas Hasenack: Approve
- Canonical Server: Pending requested
- git-ubuntu developers: Pending requested
-
Diff: 426 lines (+380/-0)7 files modifieddebian/changelog (+9/-0)
debian/patches/series (+5/-0)
debian/patches/ubuntu/lp-1771662-1-util-fixing-wrong-assumption-that-PF-has-to-have-net.patch (+62/-0)
debian/patches/ubuntu/lp-1771662-2-util-Code-simplification.patch (+104/-0)
debian/patches/ubuntu/lp-1771662-3-util-Fix-for-NULL-dereference.patch (+74/-0)
debian/patches/ubuntu/lp-1771662-4-util-Fixing-invalid-error-checking-from-virPCIGetNet.patch (+35/-0)
debian/patches/ubuntu/lp-1811198-utils-Remove-arbitrary-limit-on-socket_id-core_id.patch (+91/-0)
summary: |
- Remove arbitrary limit on socket_id/core_id + virt-manager fail to start on huawei arm server |
https:/ /www.redhat. com/archives/ libvir- list/2018- August/ msg00798. html
While in most cases the values are going to be much
smaller than our arbitrary 4096 limit, there is really
no guarantee that would be the case: in fact, a few
aarch64 servers have been spotted in the wild with
core_id as high as 6216.
Take advantage of virBitmap's ability to automatically
alter its size at runtime to accomodate such values.