Public bug reported:
This is more of a libvirt (or maybe even lower down in the kernel) bug,
but the consequence of $topic's reporting is that if libvirt CPU power
management is enabled, we mess up our NUMA accounting because we have
the wrong socket for some/all of our dedicated CPUs, depending o
Public bug reported:
Description
===
libvirt CPU power management does not support live migration
Steps to reproduce
==
1. Turn on libvirt CPU power management
2. Boot an instance with hw:cpu_policy=dedicated
3. Live migrate the instance
Expected result
===
Li
Public bug reported:
Description
===
libvirt CPU power management does not handle `isolate` emulator thread policy.
Steps to reproduce
==
1. Turn on libvirt CPU power management
2. Boot an instance with hw:cpu_policy=dedicated and
hw:emulator_threads_policy=isolate
Expec
I'm afraid this is expected behaviour, and we call it out in our API
reference [1]:
device (Optional)
body
string
Name of the device such as, /dev/vdb. Omit or set this parameter to null
for auto-assignment, if supported. If you specify this parameter, the
device must
If I understand correctly, it looks like the hostnames of your compute
hosts has changed because of a deployment error. I understand that this
is not your fault, but renaming compute host names is essentially
forbidden, as it causes the kind of breakage that you're reporting in
this bug.
We have i
Public bug reported:
Description
===
hw_disk_bus='sata' is replaced to scsi when it is attached to addtional disk.
Steps to reproduce
==
1. set hw_disk_bus=sata in a image
$openstack image set --property hw_disk_bus=sata
$openstack image set --property hw_disk_bus=sata
This has been attempted may times in the past, without success (see
references below). At this point, I'd rather set honest expectations
that this will never get done, and close this bug.
References:
Red Hat internal BZ: https://bugzilla.redhat.com/show_bug.cgi?id=1095997
Initial upstream bluepri
Well, it was still a _mistake_, but it's been out in the wild long
enough that we can't retroactively fix it without breaking someone's
scripts, so we have to leave it as is, unfortunately :(
** Changed in: nova
Status: In Progress => Won't Fix
--
You received this bug notification becaus
Public bug reported:
Description
===
Basically $summary. For lock, migrate, and unshelve, we have decorators for
validation schema that _start_ at a certain microversion (exact microversion
varies), meaning anything below that is not checked. A client could send a
request that is only v
Sounds like you fixed your host to support virtio, and then your policy
file syntax error. Is it OK if I move this to Invalid since this is not
a Nova bug?
** Changed in: nova
Status: New => Invalid
--
You received this bug notification because you are a member of Yahoo!
Engineering Team,
Cold migration assumes that the source compute host is up (and I think
has checks in the code for that). A cold migration will fail if the
source compute host is actually down and not disabled. The correct thing
to do in that case is an evacuation.
** Changed in: nova
Status: New => Invalid
Heya,
I think there's some confusion around expectations for evacuations.
Evacuations _must_ be done with the source compute fenced, and brought
back online by a human in a controlled manner to ensure evacuated
instances are destroyed properly. Any monitoring software that initiates
evacuations w
Public bug reported:
Disclaimer: I haven't reproduced this in a functional test, but based on
the traceback that I gathered from a real environment as well as the
fact that the proposed fix actually fixes this, I think my theory is
correct
Description
===
`socket` PCI NUMA policy doesn't
Hello, thanks for the report.
This falls outside the scope of a Nova bug. It's either a support
request, or a deployment question. You can try asking on the openstack-
discuss mailing list [1].
[1] https://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-
discuss
** Changed in: nova
Hello, thanks for the bug report!
IIRC the host oom killer runs per NUMA node, so in this case this sounds
like expected behaviour: node0 is out of memory (32 + 64 is bigger than
64).
** Changed in: nova
Status: New => Invalid
--
You received this bug notification because you are a membe
While Nova indeed exposes the MTU in our metadata, our source of truth
for that information is Neutron, via the `mtu` field on the Neutron
network.
[As an aside, we've had a long standing issue wherein Neutron allows the
MTU to be mutable, but there's no real support for changing the MTU
within No
Public bug reported:
Description
===
Using the libvirt driver and a host OS that uses cgroups v2 (RHEL 9,
Ubuntu Jammy), an instance with more than 16 CPUs cannot be booted.
Steps to reproduce
==
1. Boot an instance with 10 (or more) CPUs on RHEL 9 or Ubuntu Jammy
using
Public bug reported:
Description
===
When detaching an interface that was auto-created (for example, with
networks='auto' at server boot time), Nova deletes the port.
Steps to reproduce
==
1. Boot a server with an auto-created Neutron port.
2. Detach the port using the No
Hello, thanks for the bug report!
I think given how inconsistently reproducible the issue was, we can
blame this on an environmental issue as opposed to a Nova bug. I'm not
sure there's anything further to be done at this time.
** Changed in: nova
Status: New => Invalid
--
You received t
Hi Yusuf,
Yes, this is expected. The exact quantity of ping loss will depend on
the network backend (OVS in your case), how busy/loaded the VM is, the
available network bandwidth for libvirt to copy the VM memory, as well
as whether autoconverge and/or post-copy is in use.
The following is an ove
Public bug reported:
This is really a continuation of bug 1888395.
Steps to reproduce
==
Have Neutron without multiple port bindings extension.
Boot an instance with network interfaces.
Live migrate, and fail the live migration somehow (easy to do in func tests)
Expected result
=
Public bug reported:
$subject
First noticed in internal Red Hat CI of OSP 17 (based on
stable/wallaby), reproduced in upstream DNM patch [2].
tl;dr is - Nova waits for a "bind-time" external event from Neutron when
it updates the port binding back to the original host during the resize
revert, b
Public bug reported:
Description
===
We don't update instance.request_spec when we resize to a new flavor. If the
flavor's extra specs change, this is not reflected in instance.request_spec.
This goes generally unnoticed when *adding* stuff to the instance - like
PCI devices, for example
It's a valid bug, but as ports are Neutron's responsibility, I'm not
sure what can be done in this case. Neutron is free to delete a port
without checking anything about the instance it's attached to. Perhaps
this can be changed to the Neutron component, to see if folks there have
an idea?
** Proj
Public bug reported:
When running `tox -e pep8` (either in CI or locally), there are mypy
failures on nova/crypto.py
In CI, it looks like (from
https://zuul.opendev.org/t/openstack/build/1e32efc260b84a94b6a9cfeba3c80976)
pep8 start: run-test
pep8 run-test: commands[0] | bash tools/mypywrap.sh
s
/blob/v0.26.0/eventlet/green/socket.py#L20
[3] https://review.opendev.org/#/c/626952/
[4] https://bugzilla.redhat.com/show_bug.cgi?id=1860818
** Affects: nova
Importance: Undecided
Assignee: Artom Lifshitz (notartom)
Status: In Progress
--
You received this bug notification because
> when/how Nova decides that a compute node is really down there.
The compute services update their status in the database every
`report_interval` seconds [1]. If a service doesn't report after
`service_down_time`, it's considered down.
So if a compute service goes down, it'll be at most `service
Hi, thanks for the bug report!
This looks like an issue with RDO packaging than Nova itself [1]. RDO is
meant to be be installed either via TripleO [2], or you can use their
CloudSIG GA repo directly [3]. If this doesn't work for you, you can
always file a bug with RDO directly [4].
[1] https://w
d on OSP13/queens
[1]
[1] https://bugzilla.redhat.com/show_bug.cgi?id=1818829
** Affects: nova
Importance: Medium
Assignee: Artom Lifshitz (notartom)
Status: In Progress
** Tags: live-migration
--
You received this bug notification because you are a member of Yahoo!
E
Closing, as per comment #3.
** Also affects: nova/pike
Importance: Undecided
Status: New
** Changed in: nova/pike
Status: New => Fix Released
** Changed in: nova
Status: New => Fix Released
--
You received this bug notification because you are a member of Yahoo!
Enginee
I'm going to say the same thing as bug 1801733 - this is super nifty and
interesting, but realistically is not a concern and will most likely
never get addressed.
** Changed in: nova
Status: New => Won't Fix
--
You received this bug notification because you are a member of Yahoo!
Engineer
This is definitely nifty and impressive, but realistically will never be
addressed. We recommend folks deploy with TLS on all internal services,
including the message queue used for RPC. TLS makes this kind of in-
flight RPC hacking even less of a concern in practice - because let's
face it, if som
Looks like the patch in comment #12 has addressed this from Nova's POV,
so I'm setting this as Fix Release.
** Changed in: nova
Status: New => Fix Released
--
You received this bug notification because you are a member of Yahoo!
Engineering Team, which is subscribed to OpenStack Compute (
Public bug reported:
In pre-Train times, when live migrating an instance without a CPU policy
(and therefore without a NUMA topology) to a dest with a vcpu_pin_set,
or to a dest with a vcpu_pin_set different from the source, the
instance's CPU pinning information was not updated.
Now that CPU res
You mean the VNC server(s) that are created on the compute hosts for
their instances? Those are not supposed to be publically accessible.
Access to those is done via the consoles API [1] which provides an
authentication token to the client. The client the connects to the
publically-facing console p
Hello, thanks for the feature request.
Unfortunately, it is very unlikely that someone will work on it (unless
you yourself chose to do so), so I'm going to close this bug in order to
set realistic expectations. If you want to work on this yourself, please
file a blueprint.
** Changed in: nova
Based on comments #31 and #30, we consider this as 'Fix released' in
Train.
** Changed in: nova
Status: In Progress => Fix Released
--
You received this bug notification because you are a member of Yahoo!
Engineering Team, which is subscribed to OpenStack Compute (nova).
https://bugs.laun
;shelving'}
2019-05-28 03:18:48.523 26 ERROR nova.api.openstack.wsgi
Additional Info
===
This is obviously minor, as the difference between a 500 and a 409 is
purely semantic, but we're being told this is an SLA thing. An SLA
defines 5xx as being "down", while 4xx is u
Couple of things here:
You're not using hw:cpu_policy=dedicated, which means multiple instances
vCPUS can be pinned to the same host CPUs. This explains why they're all
landing on the same NUMA node.
You're not setting hw:mem_page_size, which means Nova does not do per-
NUMA-cell accounting of me
Public bug reported:
Description
===
Evacuation takes a long time if the destination host a large number of
network interfaces.
Steps to reproduce
==
1. Have a host down, or force it down.
2. Evacuate instances to a host with a large number of network
interfaces.
Expec
Public bug reported:
Description
===
If you boot an instance with pinned CPUs (for example by using the
'dedicated' CPU policy), change the vcpu_pin_set option on its compute
host, then attempt to delete the instance, it will ERROR out instead of
deleting successfully. Subsequent delete a
Public bug reported:
Description
===
When deleting an instance that hasn't finished building, we'll sometimes
get a 409 from placement as such:
Failed to delete allocations for consumer 6494d4d3-013e-478f-
9ac1-37ca7a67b776. Error: {"errors": [{"status": 409, "title":
"Conflict", "detail
Public bug reported:
test_server_connectivity_cold_migration_revert has started failing now
that we've un-skipped it [1]. It appears as though Nova is doing
everything right in terms of external events and VIF plugging (see
analysis on PS6 of [2]), so the thinking is that it's something with the
N
Public bug reported:
Description
===
Setting a device role tag on a PF interface has no effect on metadata -
IOW, the PF and its tag doesn't appear in the device metadata at all.
Steps to reproduce
==
1. Create a PF port:
openstack port show 6dd3b82f-ce2f-44dd-acd0-62
I don't think this is Nova's responsibility - as long as the new block
device shows up with `lsblk`, it's up to the guest OS/tooling to mount
it (or not). I've changed the component to cloud-init in case there's
something actionable for them in this bug. Thanks!
** Project changed: nova => cloud-i
t.com/show_bug.cgi?id=1678681
[2] https://review.opendev.org/#/c/660782/
[3] https://review.opendev.org/#/c/653498/
** Affects: nova
Importance: Undecided
Assignee: Artom Lifshitz (notartom)
Status: In Progress
--
You received this bug notification because you are a member of
Public bug reported:
Description
===
Setting [libvirt]/hw_machine_type=x86_64=q35 in nova-cpu.conf results in
any instances booted with a config drive to fail to spawn.
Steps to reproduce
==
1. Set [libvirt]/hw_machine_type=x86_64=q35 in nova-cpu.conf
2. Boot an instance
Public bug reported:
Description
===
Since CVE-2017-16239, we call to the scheduler when doing a rebuild with
a new image. If the scheduler refuses a rebuild because a filter forbids
the new image on the instance's host (for example, IsolatedHostsFilter),
at first there was no indication
Public bug reported:
Description
===
Since CVE-2017-16239, we now go through the scheduler for rebuilds. If
the scheduler refuses a rebuild with a new image because of filter
constraints (for example IsolatedHostsFilter), the instance's imageref
is set to the new image and never rolled ba
Public bug reported:
Description
===
When booting an instance with a PF passed through, device tags applied
to the PF do not appear in the metadata.
Steps to reproduce
==
Create a PF neutron port:
$ neutron port-create sriov --binding:vnic-type direct-physical --name
pf
Public bug reported:
Description
===
In Newton and earlier, 'nova usage' returns 500 as soon as the first row
in instance_extra that's marked deleted is archived by 'nova-manage db
archive_deleted_rows'.
Steps to reproduce
==
1. Create a few VMs:
$ for i in `seq 1 5`
Public bug reported:
Description
===
If an instance has had an attached volume volume-updated twice in a
"round-trip" - ie, volume-update $vol1 $vol2, then volume-update $vol2
$vol1 - it cannot be live-migrated.
Steps to reproduce
==
1. Create two iscsi volumes.
# cin
I think this is actually a valid bug in openstackclient. Specifically,
the following use case *should* work:
$ export OS_COMPUTE_API_VERSION=2.latest
$ openstack server create --flavor m1.tiny --image cirros-0.3.4-x86_64
--key-name stack --security-group default --nic net-
id=059a48f5-fc8e-4721
*** This bug is a duplicate of bug 1535918 ***
https://bugs.launchpad.net/bugs/1535918
I spent some time adding LOG.debug statements all over the place while
trying to reproduce this bug, and then read bug 1535918. They're the
same bug.
To summarise:
a. When spawning a VM, the libvirt driver
Public bug reported:
Description:
With Neutron, evacuating an instance results in a
VirtualInterfaceCreateException and the evacuation fails.
Steps to reproduce:
1. Boot a VM with a Neutron NIC.
2. Cause the underlying host to be down, for example by stopping the compute
service.
3. Evacuate t
Public bug reported:
When booting instance with a block_device_mapping with
source=blank,destination=local,tag=foo no entry is created in the
block_device_mapping table and the volume is not tagged. While the
former is existing normal behaviour, the latter is definitely a bug.
** Affects: nova
4. Boot a new instance with the same fixed IP.
This has been reported against Icehouse and has been reproduced in
master, and is therefore presumably present in all versions in-between.
** Affects: nova
Importance: Undecided
Assignee: Artom Lifshitz (notartom)
Status: New
**
confuse someone debugging an
unrelated issue.
** Affects: nova
Importance: Undecided
Assignee: Artom Lifshitz (notartom)
Status: In Progress
--
You received this bug notification because you are a member of Yahoo!
Engineering Team, which is subscribed to OpenStack Compute (n
Public bug reported:
When GETting os-volumes, the returned list of volumes respects the
osapi_max_limit configuration parameter but does not provide a link to
the next element in the list. For example, with two volumes configured
and osapi_max_limit set to 1, GETting volumes results in the followi
Public bug reported:
With osapi_max_limit set to 1 for example, querying a list of images
from Nova returns all the available images, ignoring the limit set in
the configuration.
** Affects: nova
Importance: Undecided
Status: New
--
You received this bug notification because you a
, and now it's SHA256.
** Affects: nova
Importance: Undecided
Assignee: Artom Lifshitz (notartom)
Status: New
** Changed in: nova
Assignee: (unassigned) => Artom Lifshitz (notartom)
--
You received this bug notification because you are a member of Yahoo!
Eng
Public bug reported:
get_pagination_params's [1] docstring states that "If 'limit' is not
specified, 0, or > max_limit, we default to max_limit." In practice, the
code only sets params['limit'] if 'limit' is present in the GET request.
It appears as though get_pagination_params should be a private
62 matches
Mail list logo
