[prev in list] [next in list] [prev in thread] [next in thread] 

List:       openvz-users
Subject:    Re: [Users] vzmigrate --online used to work, now fails with "CPU mismatch"
From:       jjs - mainphrame <jjs () mainphrame ! com>
Date:       2019-03-23 19:21:55
Message-ID: CABkzw7x83DRETYuRpMMW_LEXHySOpaoqsMvcJQ9Nc08YPPXQsQ () mail ! gmail ! com
[Download RAW message or body]

[Attachment #2 (multipart/alternative)]


Thanks, patched my local copies.

Jake

On Fri, Mar 22, 2019 at 4:29 PM Kirill Kolyshkin <kolyshkin@gmail.com>
wrote:

> On Fri, 22 Mar 2019 at 13:16, Konstantin Khorenko <khorenko@virtuozzo.com>
> wrote:
>
>> On 03/22/2019 07:51 PM, jjs - mainphrame wrote:
>>
>> However, one is an Intel CPU, the other is AMD. Live migration of
>> containers between them had been working, for about 3 years, but now it
>> balks at "CPUs mismatch".
>>
>> You know, you are very lucky. We do face issues from time to time when
>> processes die after online migration and
>> the root cause appears in cpus difference.
>>
>> I wonder, is there some way to override the paranoia? Ideally, an admin
>> could say "yes, I understand the CPUs aren't identical, but do it anyway"
>>
>> Here you are:
>>
>> # man vzmigrate
>>
>>        -f,
>> --nodeps[=[all][,cpu_check][,disk_space][,technologies][,license][,rate][,bindmount][,tem-
>>        plate_area_sync][,kernel_modules]]
>>               Ignore  an  absence of required package sets on destination
>> node.  To prevent CT against errors
>>               in filesystem due to absent template files, it will not be
>> started on  destination  node  after
>>               migration and must be started manually.
>>               Additional parameters:
>>               all - as is -f.
>>               cpu_check - to pass check of the cpu capabilities.
>>
>
> That's a pity that all this is written in such incomprehensible way...
> but it's Friday and so I took some time to fix this. Please see the
> attached patch.
>
>
>> --
>> Best regards,
>>
>> Konstantin Khorenko,
>> Virtuozzo Linux Kernel Team
>>
>>
>> Jake
>>
>> On Fri, Mar 22, 2019 at 9:08 AM jjs - mainphrame <jjs@mainphrame.com>
>> wrote:
>>
>>> The output on both hosts is "x86_64"
>>>
>>> Jake
>>>
>>> On Fri, Mar 22, 2019 at 1:32 AM Narcis Garcia <informatica@actiu.net>
>>> wrote:
>>>
>>>> What is the output of this command in both origin and destination hosts?
>>>> $ uname -m
>>>>
>>>>
>>>> El 21/3/19 a les 23:27, jjs - mainphrame ha escrit:
>>>> > Greetings -
>>>> >
>>>> > vzmigrate --online always worked reliably on my 2 openvz 7 servers,
>>>> but
>>>> > nowadays, vzmigrate fails, for all containers, every time.
>>>> >
>>>> > ((CPUs mismatch))) -
>>>> >
>>>> > Apologies if I missed a memo, but why has that only now become an
>>>> issue?
>>>> >
>>>> > [root@annie ~]# vzmigrate hachi 1989 --online
>>>> > Connection to destination node (hachi) is successfully established
>>>> > Moving/copying CT 1989 -> CT 1989, [], [] ...
>>>> > locking 1989
>>>> > Checking bindmounts
>>>> > Check cluster ID
>>>> > Checking keep dir for private area copy
>>>> > Check of requires kernel modules
>>>> > Checking technologies
>>>> > Checking IP addresses on destination node
>>>> > Checking RATE parameters in config
>>>> > Checking ploop format 2
>>>> > copy CT private /vz/private/1989
>>>> > Live migration stage started
>>>> > Compression is enabled
>>>> > Phaul service failed to live migrate CT
>>>> > Phaul failed to live migrate CT (/var/log/phaul.log)
>>>> > Can't move/copy CT 1989 -> CT 1989, [], [] : Phaul failed to live
>>>> > migrate CT (/var/log/phaul.log)
>>>> > unlocking 1989
>>>> > [root@annie ~]# tail /var/log/phaul.log
>>>> >     load_entry_point('phaul==0.1', 'console_scripts', 'p.haul')()
>>>> >   File "/usr/lib/python2.7/site-packages/phaul/shell/phaul_client.py",
>>>> > line 49, in main
>>>> >     worker.start_migration()
>>>> >   File "/usr/lib/python2.7/site-packages/phaul/iters.py", line 161, in
>>>> > start_migration
>>>> >     self.__start_live_migration()
>>>> >   File "/usr/lib/python2.7/site-packages/phaul/iters.py", line 175, in
>>>> > __start_live_migration
>>>> >     self.__validate_cpu()
>>>> >   File "/usr/lib/python2.7/site-packages/phaul/iters.py", line 114, in
>>>> > __validate_cpu
>>>> >     raise Exception("CPUs mismatch")
>>>> > Exception: CPUs mismatch
>>>> > [root@annie ~]#
>>>> >
>>>> > Regards,
>>>> >
>>>>
>>> _______________________________________________
> Users mailing list
> Users@openvz.org
> https://lists.openvz.org/mailman/listinfo/users
>

[Attachment #5 (text/html)]

<div dir="ltr">Thanks, patched my local copies.  \
<div><br></div><div>Jake</div></div><br><div class="gmail_quote"><div dir="ltr" \
class="gmail_attr">On Fri, Mar 22, 2019 at 4:29 PM Kirill Kolyshkin &lt;<a \
href="mailto:kolyshkin@gmail.com">kolyshkin@gmail.com</a>&gt; \
wrote:<br></div><blockquote class="gmail_quote" style="margin:0px 0px 0px \
0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex"><div dir="ltr"><div \
dir="ltr"><div dir="ltr"><div dir="ltr">On Fri, 22 Mar 2019 at 13:16, Konstantin \
Khorenko &lt;<a href="mailto:khorenko@virtuozzo.com" \
target="_blank">khorenko@virtuozzo.com</a>&gt; wrote:<br></div><div \
class="gmail_quote"><blockquote class="gmail_quote" style="margin:0px 0px 0px \
0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex">



<div bgcolor="#FFFFFF">
On 03/22/2019 07:51 PM, jjs - mainphrame wrote:<br>
<blockquote type="cite">
<div dir="ltr">However, one is an Intel CPU, the other is AMD. Live migration of \
containers between them had been working, for about 3 years, but now it balks at \
&quot;CPUs mismatch&quot;.<br> </div>
</blockquote>
You know, you are very lucky. We do face issues from time to time when processes die \
after online migration and<br> the root cause appears in cpus difference.<br>
<br>
<blockquote type="cite">
<div dir="ltr">I wonder, is there some way to override the paranoia? Ideally, an \
admin could say &quot;yes, I understand the CPUs aren&#39;t identical, but do it \
anyway&quot;<br> </div>
</blockquote>
Here you are:<br>
<br>
# man vzmigrate<br>
<br>
             -f,                 \
--nodeps[=[all][,cpu_check][,disk_space][,technologies][,license][,rate][,bindmount][,tem-<br>
  plate_area_sync][,kernel_modules]]<br>
                           Ignore   an   absence of required package sets on \
                destination node.   To prevent CT against errors<br>
                           in filesystem due to absent template files, it will not be \
started on   destination   node   after<br>  migration and must be started \
manually.<br>  Additional parameters:<br>
                           all - as is -f.<br>
                           cpu_check - to pass check of the cpu \
capabilities.<br></div></blockquote><div><br></div><div>That&#39;s a pity that all \
this is written in such incomprehensible way...</div><div>but it&#39;s Friday and so \
I took some time to fix this. Please see the attached \
patch.</div><div><div><br></div></div><blockquote class="gmail_quote" \
style="margin:0px 0px 0px 0.8ex;border-left:1px solid \
rgb(204,204,204);padding-left:1ex"><div bgcolor="#FFFFFF"> <br>
--<br>
Best regards,<br>
<br>
Konstantin Khorenko,<br>
Virtuozzo Linux Kernel Team<br>
<br>
<blockquote type="cite">
<div dir="ltr"><br>
Jake</div>
<br>
<div class="gmail_quote">
<div dir="ltr" class="gmail_attr">On Fri, Mar 22, 2019 at 9:08 AM jjs - mainphrame \
&lt;<a href="mailto:jjs@mainphrame.com" target="_blank">jjs@mainphrame.com</a>&gt; \
wrote:<br> </div>
<blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left:1px solid \
rgb(204,204,204);padding-left:1ex"> <div dir="ltr">
<div dir="ltr">The output on both hosts is &quot;x86_64&quot;</div>
<div dir="ltr"><br>
</div>
<div>Jake</div>
</div>
<br>
<div class="gmail_quote">
<div dir="ltr" class="gmail_attr">On Fri, Mar 22, 2019 at 1:32 AM Narcis Garcia \
&lt;<a href="mailto:informatica@actiu.net" \
target="_blank">informatica@actiu.net</a>&gt; wrote:<br> </div>
<blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left:1px solid \
rgb(204,204,204);padding-left:1ex"> What is the output of this command in both origin \
and destination hosts?<br> $ uname -m<br>
<br>
<br>
El 21/3/19 a les 23:27, jjs - mainphrame ha escrit:<br>
&gt; Greetings -  <br>
&gt; <br>
&gt; vzmigrate --online always worked reliably on my 2 openvz 7 servers, but<br>
&gt; nowadays, vzmigrate fails, for all containers, every time.  <br>
&gt; <br>
&gt; ((CPUs mismatch))) -<br>
&gt; <br>
&gt; Apologies if I missed a memo, but why has that only now become an issue?<br>
&gt; <br>
&gt; [root@annie ~]# vzmigrate hachi 1989 --online<br>
&gt; Connection to destination node (hachi) is successfully established<br>
&gt; Moving/copying CT 1989 -&gt; CT 1989, [], [] ...<br>
&gt; locking 1989<br>
&gt; Checking bindmounts<br>
&gt; Check cluster ID<br>
&gt; Checking keep dir for private area copy<br>
&gt; Check of requires kernel modules<br>
&gt; Checking technologies<br>
&gt; Checking IP addresses on destination node<br>
&gt; Checking RATE parameters in config<br>
&gt; Checking ploop format 2<br>
&gt; copy CT private /vz/private/1989<br>
&gt; Live migration stage started<br>
&gt; Compression is enabled<br>
&gt; Phaul service failed to live migrate CT<br>
&gt; Phaul failed to live migrate CT (/var/log/phaul.log)<br>
&gt; Can&#39;t move/copy CT 1989 -&gt; CT 1989, [], [] : Phaul failed to live<br>
&gt; migrate CT (/var/log/phaul.log)<br>
&gt; unlocking 1989<br>
&gt; [root@annie ~]# tail /var/log/phaul.log<br>
&gt;       load_entry_point(&#39;phaul==0.1&#39;, &#39;console_scripts&#39;, \
&#39;p.haul&#39;)()<br> &gt;    File \
&quot;/usr/lib/python2.7/site-packages/phaul/shell/phaul_client.py&quot;,<br> &gt; \
line 49, in main<br> &gt;       worker.start_migration()<br>
&gt;    File &quot;/usr/lib/python2.7/site-packages/phaul/iters.py&quot;, line 161, \
in<br> &gt; start_migration<br>
&gt;       self.__start_live_migration()<br>
&gt;    File &quot;/usr/lib/python2.7/site-packages/phaul/iters.py&quot;, line 175, \
in<br> &gt; __start_live_migration<br>
&gt;       self.__validate_cpu()<br>
&gt;    File &quot;/usr/lib/python2.7/site-packages/phaul/iters.py&quot;, line 114, \
in<br> &gt; __validate_cpu<br>
&gt;       raise Exception(&quot;CPUs mismatch&quot;)<br>
&gt; Exception: CPUs mismatch<br>
&gt; [root@annie ~]#  <br>
&gt; <br>
&gt; Regards,<br>
&gt;  <br></blockquote></div></blockquote></div></blockquote></div>
</blockquote></div></div></div></div>
_______________________________________________<br>
Users mailing list<br>
<a href="mailto:Users@openvz.org" target="_blank">Users@openvz.org</a><br>
<a href="https://lists.openvz.org/mailman/listinfo/users" rel="noreferrer" \
target="_blank">https://lists.openvz.org/mailman/listinfo/users</a><br> \
</blockquote></div>



_______________________________________________
Users mailing list
Users@openvz.org
https://lists.openvz.org/mailman/listinfo/users


[prev in list] [next in list] [prev in thread] [next in thread] 

Configure | About | News | Add a list | Sponsored by KoreLogic