错误：实例热迁移到主机"compute06"失败 Failed to compute_task_migrate_server: Unacce

admin

2022-06-21 14:47:33.988 25 WARNING oslo_config.cfg [req-b990f037-8329-4f82-aba0-2ad6df0942d8 ef9faa1589a945ec9764b05c1c433b57 6f0124196ea74eb79fbcf370add3ca7e - default default] Option "os_region_name" from group "placement" is deprecated. Use option "region-name" from group "placement".
* p4 U- G1 t7 j9 h6 i2022-06-21 14:47:36.470 25 WARNING nova.scheduler.utils [req-b990f037-8329-4f82-aba0-2ad6df0942d8 ef9faa1589a945ec9764b05c1c433b57 6f0124196ea74eb79fbcf370add3ca7e - default default] Failed to compute_task_migrate_server: Unacceptable CPU info: CPU doesn't have compatibility.
4 t0 X  x, c) h  h/ e1 N
- g, H/ b0 B' y0 h0
5 Z) s: V3 \# `/ p& \
6 G+ A5 c! e* F  N. HRefer to http://libvirt.org/html/libvirt- ... virCPUCompareResult
Traceback (most recent call last):

  File "/usr/lib/python2.7/site-packages/oslo_messaging/rpc/server.py", line 166, in _process_incoming
    res = self.dispatcher.dispatch(message)
2 I0 ^9 o4 P( t3 E' g- |
! O8 D5 Q& o6 |- [* f5 ~- K  File "/usr/lib/python2.7/site-packages/oslo_messaging/rpc/dispatcher.py", line 220, in dispatch
$ ~3 u0 S$ c5 B4 U0 t    return self._do_dispatch(endpoint, method, ctxt, args)
  W0 U7 E) p2 s* ]" U
0 {# B; V" p% n  File "/usr/lib/python2.7/site-packages/oslo_messaging/rpc/dispatcher.py", line 190, in _do_dispatch
    result = func(ctxt, **new_args)
2 G0 t5 n" z5 T. o9 ~+ d. [/ z9 v
0 \! W! `7 q2 J) m7 b  File "/usr/lib/python2.7/site-packages/nova/exception_wrapper.py", line 76, in wrapped
* R) j; y3 Q  t7 z' [    function_name, call_dict, binary)

$ N  t, G) o, l& m: f9 A' K' i  File "/usr/lib/python2.7/site-packages/oslo_utils/excutils.py", line 220, in __exit__
, ~" c" g5 B% ?    self.force_reraise()

  File "/usr/lib/python2.7/site-packages/oslo_utils/excutils.py", line 196, in force_reraise
; s+ C* a+ G) n3 x    six.reraise(self.type_, self.value, self.tb)

# M: U( G6 d6 m" d  File "/usr/lib/python2.7/site-packages/nova/exception_wrapper.py", line 67, in wrapped
    return f(self, context, *args, **kw)
9 t$ c8 A5 v3 U, f4 n. @4 d4 i
0 q6 g2 B5 }0 S* J# C  File "/usr/lib/python2.7/site-packages/nova/compute/utils.py", line 1000, in decorated_function
    return function(self, context, *args, **kwargs)

/ s4 l: ?7 ^, s3 W" `; B1 a: }/ |  File "/usr/lib/python2.7/site-packages/nova/compute/manager.py", line 215, in decorated_function
    kwargs['instance'], e, sys.exc_info())
4 X. X6 E0 i0 a7 o
  File "/usr/lib/python2.7/site-packages/oslo_utils/excutils.py", line 220, in __exit__
' Q5 h6 O0 x8 p: Y) w7 y) {    self.force_reraise()

  File "/usr/lib/python2.7/site-packages/oslo_utils/excutils.py", line 196, in force_reraise
    six.reraise(self.type_, self.value, self.tb)

+ f; A! H) M5 ?- m/ g$ c5 V  File "/usr/lib/python2.7/site-packages/nova/compute/manager.py", line 203, in decorated_function
& G- |9 }' [+ z* ^, p5 P4 g. i2 d    return function(self, context, *args, **kwargs)
" u8 t: d! a" R9 S8 K: w0 Z$ A
  File "/usr/lib/python2.7/site-packages/nova/compute/manager.py", line 5971, in check_can_live_migrate_destination
4 N$ J' s1 F9 p  i  E# e5 v    disk_over_commit)

  File "/usr/lib/python2.7/site-packages/nova/compute/manager.py", line 5982, in _do_check_can_live_migrate_destination
6 z# r% X1 h8 m    block_migration, disk_over_commit)
$ K8 j3 B# a1 M
. n, S- w3 ^" g3 V/ E  File "/usr/lib/python2.7/site-packages/nova/virt/libvirt/driver.py", line 6570, in check_can_live_migrate_destination
    self._compare_cpu(None, source_cpu_info, instance)

( s+ r& h6 @  L2 C5 G  File "/usr/lib/python2.7/site-packages/nova/virt/libvirt/driver.py", line 6841, in _compare_cpu
* ^% Y' |+ U, O' u8 v* t! }    raise exception.InvalidCPUInfo(reason=m % {'ret': ret, 'u': u})
* y8 b* ~: b0 b8 U" X
6 w1 a7 _1 X, KInvalidCPUInfo: Unacceptable CPU info: CPU doesn't have compatibility.

0
8 m* }$ U* e/ {
Refer to http://libvirt.org/html/libvirt- ... virCPUCompareResult
: y; ~' M' R) C0 w6 `: InvalidCPUInfo_Remote: Unacceptable CPU info: CPU doesn't have compatibility.
2022-06-21 14:47:36.472 25 WARNING nova.scheduler.utils [req-b990f037-8329-4f82-aba0-2ad6df0942d8 ef9faa1589a945ec9764b05c1c433b57 6f0124196ea74eb79fbcf370add3ca7e - default default] [instance: 4e781528-c494-45fe-8c67-15763fd5650c] Setting instance to ACTIVE state.: InvalidCPUInfo_Remote: Unacceptable CPU info: CPU doesn't have compatibility.

admin

Live migration fails with 'Unacceptable CPU info: CPU doesn't have compatibility' between non-rebooted and rebooted compute nodes after in-place OpenStack Platform major upgrade
) W' K% Z' a, p9 E% f4 p+ x SOLUTION UNVERIFIED - 已更新 2018年四月23日15:25 - English
环境
Red Hat OpenStack Platform after major upgrade.
Red Hat Enterprise Linux (RHEL) 7 on overcloud nodes.
microcode_ctl-2.1-22.5.el7_4.x86_64 and linux-firmware-20170606-58.gitc990aae.el7_4.noarch installed as part of the upgrade.
Non-rebooted source compute node is running microcode version 59 and rebooted target is running microcode version 58.
7 N2 Q. F+ r5 k4 u" xThe IBPB CPU model is in use on the source node but not the target node.
问题
After an in-place OpenStack Platform major upgrade, live migration fails with the following error when the source node has not been rebooted and the target node has:
/ i( o" g  t6 ^+ ^& X/ b; k, ]
, Q: I7 M- ]" `5 F, }7 W" {Raw
Unacceptable CPU info: CPU doesn't have compatibility
决议
% |5 [+ @4 L& \; L9 j$ S2 `We recommend working with your respective silicon/hardware vendors in order to upgrade target nodes to version 59 of the CPU microcode or later (whichever is recommended by your silicon/hardware vendors).

After updating, confirm whether virsh capabilities reports that microcode version 59 or later is installed on target node and that the IBPB CPU model is in use. If this is the case, then test live migration on a non-production guest first.

Following this, complete similar microcode updates on all of the other hosts within the OSP environment.

% ]% F/ ~9 \9 F. p9 z根源
If Red Hat distributed microcode_ctl and linux-firmware packages are downgraded to microcode version 58 during a major upgrade, the microcode downgrade will only take effect on each node after a reboot.
+ i, S, C! m) P$ a6 M) y
  F4 c: q2 G" H" Z4 \& A; V; dThe latest microcode_ctl and linux-firmware packages from Red Hat do not include resolutions to CVE-2017-5715 (Spectre variant 2) exploit. Red Hat is no longer providing microcode to address Spectre variant 2, due to instabilities introduced that are causing customer systems to not boot. The latest microcode_ctl and linux-firmware packages are reverting these unstable microprocessor firmware changes to versions that were known to be stable and well tested, released prior to the Spectre/Meltdown embargo lift date on January 3rd 2018. The following article provides more detail in relation to this:

& V$ t4 b, \1 Q+ M  y+ |• What CPU microcode is available via the microcode_ctl package to mitigate CVE-2017-5715 (variant 2)?

In certain scenarios (for example if compute nodes are not rebooted during the major upgrade procedure and then target nodes are later rebooted post upgrade but source nodes are not), source nodes could be running microcode version 59 or later whereas rebooted target nodes could be running microcode version 58.

. v$ A# t* y) X# [3 g" iAs such, live migrations would then fail with the Unacceptable CPU info: CPU doesn't have compatibility error, because the target node would lack support for required CPU features.
- [. _& C5 Y% G! M
7 j, P* s* }8 c4 N4 D/ U诊断步骤
2 d9 a4 Y, g" j7 oTo check which microcode version is actually running on each node, one option would be to run virsh capabilities and review the microcode version= parameter within the output. For example:

+ }$ U6 N# H8 w5 FRaw
<microcode version='58'/>
. o8 b/ w0 i: R: yAlternatively, /proc/cpuinfo provides the microcode version in hexadecimal format. For example:

Raw
0 q8 T7 U3 @7 E! P+ G2 T# }$ grep -is '^microcode' /proc/cpuinfo | head -1
* L% B* s4 B0 |" a$ `microcode   : 0x3a
We can see that the above example reports microcode version 58 is running:
3 z7 g7 Y' K9 n0 P5 j- B, a
Raw
$ echo "ibase=16;3A" | bc
58

admin

Live Migration Fails With Error: "Unacceptable CPU Info: CPU Doesn't Have Compatibility.
Problem
6 W3 |' o0 g$ z# uAttempts to perform a live migration of an instance fail with the following error.

Unacceptable CPU info: CPU doesn't have compatibility.
: e/ c) W: w" w8 A7 P0 T% {Copy
3 N# `. W6 }1 w' @, ]' ?Environment
  Y6 @& [& s; y2 XPlatform9 Managed OpenStack - v3.6.0 and Higher
* ?7 C/ h* D1 R, u0 [Cause
) P+ F" f/ t& tThis error presents itself when CPU model/architecture differ between the source host and destination host even slightly.
& M3 U! y0 \$ K
Resolution
Run the following command on the host where the VM currently resides to create an XML with the full CPU description.
# virsh capabilities > virsh-caps-[originHostname]-full.xml
/ c4 J. r% P! I' l! s9 o! PCopy
Edit the XML file you've generated deleting everything above "[cpu]" and everything below "[/cpu]".
! q7 L) T1 |+ R8 Q7 }3 N0 e8 e, EMove the file to the destination host and run the following command which will tell you if there is a legitimate mismatch and on which side it lies.
/ {# ~  k5 V# N# virsh cpu-compare virsh-caps-[originHostname]-full.xml
' A: \0 y+ i( m5 ICopy
To determine the specific features that must be masked out, run the following command.
# virsh cpu-baseline virsh-caps-[originHostname]-full.xml
Copy
Mask out the features that do not appear in the output of the command in Step #4 and attempt to migrate the VM again.