[ovirt-users] when engine broken

[张 余歌]

recently i confused about how i can recovery engine once engine process failed
or broken for some reason.

enviroment:allinone install ovirt.3.5.x
engine installed on local host whithout any protection.i can do some engine 
backup by using command.

I refer to some way to recovery engine about disater recovery .but it seem 
maybe my stage is wrong or other issues,whould u give me some way to recover 
engine,and find my lost vm？？
thanks a lot

获取 Outlook for Android
___
Users mailing list
Users@ovirt.org
http://lists.ovirt.org/mailman/listinfo/users

Re: [ovirt-users] Local and Shared storage in same datacenter

[Mike]

Op 30-10-2016 om 12:50 schreef Yaniv Kaul:

Perhaps, if there are no HA requirements, those VMs with local
domain
needs can be in their own DC, a local one? If it's just 5%,
shouldn't be
much of an effort?

Can a host be in 2 datacenters?


No, I suggested to dedicate a host (or several) with their own local DC
for those 5%.
Y.
I have 5 hosts and dedicating 2 hosts for 5% seems a gigantic waste if 
you ask me.

So I cannot do that.
___
Users mailing list
Users@ovirt.org
http://lists.ovirt.org/mailman/listinfo/users

Re: [ovirt-users] 100% disk utilization on Hosted Engine

[Anantha Raghava]

Hello Didi,

Apart form oVirt-engine, nothing else is running on the server. Engine 
is running on CentOS 7.2 minimal installation. We have 2 hosts, 1 iSCSI 
storage, about 19 VMs and 5 Networks. In this setup oVirt engine is 
running in a separate VM created on VMWare ESXi. In 20 odd days, it 
consumed around 70GB out of 100 GB!!!


In another setup, we have only one host, hosted engine, 9 VMs, 2 
Networks, 1 FC Storage. In about 30 odd days, 100% of available 100 GB 
consumed!! It is in this setup CentOS is not allowing us to loginto 
engine VM as no space to create any temp files on the storage. We have 
ISO Domain created on this engine VM. But that's only about 3 GB in size.


Both setup are not exposed to internet. Hence, any malware is ruled out.

We have observed, the logs are also growing and we have not found that 
logs are not rotating. All logs right from day one are kept intact. 
Nothing in the user files. Because no user is allowed to use this VM.


What else it could be? 100% consumption without even a warning, is very 
strange.


--

Thanks & Regards,


Anantha Raghava

eXza Technology Consulting & Services


Do not print this e-mail unless required. Save Paper & trees.

On Sunday 30 October 2016 07:20 PM, Yedidyah Bar David wrote:

On Sun, Oct 30, 2016 at 3:01 PM, Anantha Raghava
 wrote:

Hi,

In another instance where the engine is on a different host, we have
observed that out of 100 GB nearly 70% is consumed in just about 20 odd
days.

Is it that DWH database and the logs really take so much of space so fast?

Are you sure that DWH database is the only thing using much space? Did you
check disk space usage?

DWH db size depends on the size of your setup - number of
hosts/VMs/disks/NICs/etc.

You can also affect this by configuring granularity and history
length, check e.g.:

https://www.mail-archive.com/search?l=users@ovirt.org=subject:%22Re\%3A+\[ovirt\-users\]+Limit+DWH+database+size%22=newest

http://lists.ovirt.org/pipermail/users/2016-September/042419.html


How do we reduce the disk utilization? This will be a major issue when
compared to other hypervisors management platform.

How do we address this and arrest this super fast disc consumption?

--

Thanks & Regards,


Anantha Raghava eXza Technology Consulting & Services Do not print this
e-mail unless required. Save Paper & trees.
On Saturday 29 October 2016 08:32 AM, Anantha Raghava wrote:

Hi,

I have hit a unique problem. Using oVirt 4.0.2.7-1.el7.centos since last one
month and I have hit a unique problem. The Engine is reporting that Disk
space allocated to Hosted engine is full and is not allowing to login to
engine's console to clear some files as well. When I try to add additional
storage to Hosted engine, engine admin portal reports that as Engine VM is
not managed by oVirt Engine, it cannot extend the storage.

Now the questions are:

1. Is it that enabling advanced DWH on the engine host is the root cause for
100GB of disk getting filled in about 1 month?

2. How do we clear the disk of some unwanted files and extend the space?

3. How do we reset the DWH database to basic?


Note: CentOS 7.2 is not even allowing us to login to console to do any
action. Booting the Hosted Engine VM with Live CD etc., looks like is ruled
out in this case.
--

Thanks & Regards,


Anantha Raghava eXza Technology Consulting & Services

Do not print this e-mail unless required. Save Paper & trees.



___
Users mailing list
Users@ovirt.org
http://lists.ovirt.org/mailman/listinfo/users






___
Users mailing list
Users@ovirt.org
http://lists.ovirt.org/mailman/listinfo/users

Re: [ovirt-users] 100% disk utilization on Hosted Engine

[Yedidyah Bar David]

On Sun, Oct 30, 2016 at 3:01 PM, Anantha Raghava
 wrote:
> Hi,
>
> In another instance where the engine is on a different host, we have
> observed that out of 100 GB nearly 70% is consumed in just about 20 odd
> days.
>
> Is it that DWH database and the logs really take so much of space so fast?

Are you sure that DWH database is the only thing using much space? Did you
check disk space usage?

DWH db size depends on the size of your setup - number of
hosts/VMs/disks/NICs/etc.

You can also affect this by configuring granularity and history
length, check e.g.:

https://www.mail-archive.com/search?l=users@ovirt.org=subject:%22Re\%3A+\[ovirt\-users\]+Limit+DWH+database+size%22=newest

http://lists.ovirt.org/pipermail/users/2016-September/042419.html

> How do we reduce the disk utilization? This will be a major issue when
> compared to other hypervisors management platform.
>
> How do we address this and arrest this super fast disc consumption?
>
> --
>
> Thanks & Regards,
>
>
> Anantha Raghava eXza Technology Consulting & Services Do not print this
> e-mail unless required. Save Paper & trees.
> On Saturday 29 October 2016 08:32 AM, Anantha Raghava wrote:
>
> Hi,
>
> I have hit a unique problem. Using oVirt 4.0.2.7-1.el7.centos since last one
> month and I have hit a unique problem. The Engine is reporting that Disk
> space allocated to Hosted engine is full and is not allowing to login to
> engine's console to clear some files as well. When I try to add additional
> storage to Hosted engine, engine admin portal reports that as Engine VM is
> not managed by oVirt Engine, it cannot extend the storage.
>
> Now the questions are:
>
> 1. Is it that enabling advanced DWH on the engine host is the root cause for
> 100GB of disk getting filled in about 1 month?
>
> 2. How do we clear the disk of some unwanted files and extend the space?
>
> 3. How do we reset the DWH database to basic?
>
>
> Note: CentOS 7.2 is not even allowing us to login to console to do any
> action. Booting the Hosted Engine VM with Live CD etc., looks like is ruled
> out in this case.
> --
>
> Thanks & Regards,
>
>
> Anantha Raghava eXza Technology Consulting & Services
>
> Do not print this e-mail unless required. Save Paper & trees.
>
>
>
> ___
> Users mailing list
> Users@ovirt.org
> http://lists.ovirt.org/mailman/listinfo/users
>



-- 
Didi
___
Users mailing list
Users@ovirt.org
http://lists.ovirt.org/mailman/listinfo/users

Re: [ovirt-users] 100% disk utilization on Hosted Engine

[Anantha Raghava]

Hi,

In another instance where the engine is on a different host, we have 
observed that out of 100 GB nearly 70% is consumed in just about 20 odd 
days.


Is it that DWH database and the logs really take so much of space so 
fast? How do we reduce the disk utilization? This will be a major issue 
when compared to other hypervisors management platform.


How do we address this and arrest this super fast disc consumption?

--

Thanks & Regards,


Anantha Raghava

eXza Technology Consulting & Services

Do not print this e-mail unless required. Save Paper & trees.
On Saturday 29 October 2016 08:32 AM, Anantha Raghava wrote:


Hi,

I have hit a unique problem. Using oVirt 4.0.2.7-1.el7.centos since 
last one month and I have hit a unique problem. The Engine is 
reporting that Disk space allocated to Hosted engine is full and is 
not allowing to login to engine's console to clear some files as well. 
When I try to add additional storage to Hosted engine, engine admin 
portal reports that as Engine VM is not managed by oVirt Engine, it 
cannot extend the storage.


Now the questions are:

1. Is it that enabling advanced DWH on the engine host is the root 
cause for 100GB of disk getting filled in about 1 month?


2. How do we clear the disk of some unwanted files and extend the space?

3. How do we reset the DWH database to basic?


Note: CentOS 7.2 is not even allowing us to login to console to do any 
action. Booting the Hosted Engine VM with Live CD etc., looks like is 
ruled out in this case.

--

Thanks & Regards,


Anantha Raghava

eXza Technology Consulting & Services


Do not print this e-mail unless required. Save Paper & trees.



___
Users mailing list
Users@ovirt.org
http://lists.ovirt.org/mailman/listinfo/users

Re: [ovirt-users] Local and Shared storage in same datacenter

[Yaniv Kaul]

On Sun, Oct 30, 2016 at 12:47 PM, Mike (maillinglists) 
wrote:

> Hi Yaniv,
>
> Op 30-10-2016 om 11:35 schreef Yaniv Kaul:
>
>>
>>
>> On Sun, Oct 30, 2016 at 12:02 PM, Mike (maillinglists)
>> > wrote:
>>
>> Hi guys,
>>
>> There have been a few related questions already that I could find,
>> but I did not find anything relating to my specific use case.
>>
>> Currently it is not possible to mix local storage with shared
>> storage in the same datacenter.
>> The reason seems to be because of the storage pool manager (SPM).
>> This is a role in the datacenter provided to one specific host.
>>
>> While I understand that this makes having local storage impossible,
>> I believe there is a use case to have local storage in a shared
>> storage datacenter.
>>
>>
>> Indeed, this is one of the more appealing use cases. There are others s
>> well.
>>
>>
>> Consider the following:
>> I have a few applications that require 1 milli second latency and at
>> most 2 milli second.
>>
>>
>> For read, write or both?
>>
> Both I'm afraid.
>
>>
>>
>> That is not consistenly achievable with shared storage, to that end
>> I added flash storage to a few hypervisors.
>>
>>
>> You could have flash on your shared storage.
>>
> The problem with shared flash storage is the storage network required in
> between. I have seen shared NFS storage at 1ms latency, but that has not
> been stable and this application requires stable 1ms latencies.
>
>>
>>
>> About 5% of my servers require this and are not that resource hungry
>> to require a dedicated physical server.
>> That same 5% also has no requirement to be migrated if a host fails.
>>
>> So in short I have 5 heavy hosts running ovirt with a shared storage
>> domain on NFS for 95% of my servers.
>> All running fine, but I am now unable to run my remaining 5%.
>>
>>
>> Perhaps, if there are no HA requirements, those VMs with local domain
>> needs can be in their own DC, a local one? If it's just 5%, shouldn't be
>> much of an effort?
>>
> Can a host be in 2 datacenters?
>

No, I suggested to dedicate a host (or several) with their own local DC for
those 5%.
Y.


> I was not aware of this and I will investigate.
>
>>
>>
>> To finish up my summary I have been testing various virtualization
>> technologies, like VmWare and Hyper-V.
>> They allow such configurations as I mentioned.
>>
>> I already had some chat on irc with various guys and they suggested
>> that I put this on the mailing list, so here goes.
>>
>> My suggestion would be to evoluate from SPM to SDM.
>>
>>
>> Easier said than done... We have worked on this for quite some time,
>> it's not as easy as one might think.
>>
> I appreciate your honousty, but still hope this can be achieved.
>
>>
>>
>> SDM stands for Storage Domain Manager.
>> This would create the possibility to have all nodes in the
>> datacenter participate in the storage handling.
>> A extra benefit would be that local storage could be added.
>>
>> What do you think?
>>
>>
>> There are other use cases we think flash on the host can be used, some
>> may be of use for your use case.
>> For example, dm-cache[1].
>>
>> We are still looking at this. I think Gluster already can make use it
>> for cache, for example.
>> Y.
>>
> Gluster is an angle we are investigating, but takes time to look into.
> Thanks!
>
>>
>> [1] https://people.redhat.com/mskinner/rhug/q1.2016/dm-cache.pdf
>>
> Thanks!
>
>>
>>
>>
>> Thanks for reading.
>>
>> Kind regards,
>> Mike van Goor
>> ___
>> Users mailing list
>> Users@ovirt.org 
>> http://lists.ovirt.org/mailman/listinfo/users
>> 
>>
>>
>>
>
___
Users mailing list
Users@ovirt.org
http://lists.ovirt.org/mailman/listinfo/users

[ovirt-users] Fwd: Re: Local and Shared storage in same datacenter

[Mike]

Hi Barak,

Op 30-10-2016 om 11:19 schreef Barak Korren:

While I understand that this makes having local storage impossible, I
believe there is a use case to have local storage in a shared storage
datacenter.
Consider the following:
I have a few applications that require 1 milli second latency and at most 2
milli second.
That is not consistenly achievable with shared storage, to that end I added
flash storage to a few hypervisors.
About 5% of my servers require this and are not that resource hungry to
require a dedicated physical server.
That same 5% also has no requirement to be migrated if a host fails.



I mentioned this in another thread already, please look into the VDSM
scratchpad hook. It will allow you to attach (files on) the local SSDs
(as disks) to VMs when those VMs start up. It will perfectly meet your
use case if you only need to keep the data on the SSD while the VM is
up.



The data should not be lost if we power down the VM.
If the data is lost during a host failure we have options to restore the 
data, but do not want to do it on a weekly/monthly basis.


The scratchpad is useful for other options we are exploring, thanks!
___
Users mailing list
Users@ovirt.org
http://lists.ovirt.org/mailman/listinfo/users

Re: [ovirt-users] Local and Shared storage in same datacenter

[Mike (maillinglists)]

Hi Yaniv,

Op 30-10-2016 om 11:35 schreef Yaniv Kaul:



On Sun, Oct 30, 2016 at 12:02 PM, Mike (maillinglists)
> wrote:

Hi guys,

There have been a few related questions already that I could find,
but I did not find anything relating to my specific use case.

Currently it is not possible to mix local storage with shared
storage in the same datacenter.
The reason seems to be because of the storage pool manager (SPM).
This is a role in the datacenter provided to one specific host.

While I understand that this makes having local storage impossible,
I believe there is a use case to have local storage in a shared
storage datacenter.


Indeed, this is one of the more appealing use cases. There are others s
well.


Consider the following:
I have a few applications that require 1 milli second latency and at
most 2 milli second.


For read, write or both?

Both I'm afraid.



That is not consistenly achievable with shared storage, to that end
I added flash storage to a few hypervisors.


You could have flash on your shared storage.
The problem with shared flash storage is the storage network required in 
between. I have seen shared NFS storage at 1ms latency, but that has not 
been stable and this application requires stable 1ms latencies.



About 5% of my servers require this and are not that resource hungry
to require a dedicated physical server.
That same 5% also has no requirement to be migrated if a host fails.

So in short I have 5 heavy hosts running ovirt with a shared storage
domain on NFS for 95% of my servers.
All running fine, but I am now unable to run my remaining 5%.


Perhaps, if there are no HA requirements, those VMs with local domain
needs can be in their own DC, a local one? If it's just 5%, shouldn't be
much of an effort?

Can a host be in 2 datacenters?
I was not aware of this and I will investigate.



To finish up my summary I have been testing various virtualization
technologies, like VmWare and Hyper-V.
They allow such configurations as I mentioned.

I already had some chat on irc with various guys and they suggested
that I put this on the mailing list, so here goes.

My suggestion would be to evoluate from SPM to SDM.


Easier said than done... We have worked on this for quite some time,
it's not as easy as one might think.

I appreciate your honousty, but still hope this can be achieved.



SDM stands for Storage Domain Manager.
This would create the possibility to have all nodes in the
datacenter participate in the storage handling.
A extra benefit would be that local storage could be added.

What do you think?


There are other use cases we think flash on the host can be used, some
may be of use for your use case.
For example, dm-cache[1].

We are still looking at this. I think Gluster already can make use it
for cache, for example.
Y.
Gluster is an angle we are investigating, but takes time to look into. 
Thanks!


[1] https://people.redhat.com/mskinner/rhug/q1.2016/dm-cache.pdf

Thanks!




Thanks for reading.

Kind regards,
Mike van Goor
___
Users mailing list
Users@ovirt.org 
http://lists.ovirt.org/mailman/listinfo/users





___
Users mailing list
Users@ovirt.org
http://lists.ovirt.org/mailman/listinfo/users

Re: [ovirt-users] Local and Shared storage in same datacenter

[Yaniv Kaul]

On Sun, Oct 30, 2016 at 12:02 PM, Mike (maillinglists) 
wrote:

> Hi guys,
>
> There have been a few related questions already that I could find, but I
> did not find anything relating to my specific use case.
>
> Currently it is not possible to mix local storage with shared storage in
> the same datacenter.
> The reason seems to be because of the storage pool manager (SPM).
> This is a role in the datacenter provided to one specific host.
>
> While I understand that this makes having local storage impossible, I
> believe there is a use case to have local storage in a shared storage
> datacenter.
>

Indeed, this is one of the more appealing use cases. There are others s
well.


> Consider the following:
> I have a few applications that require 1 milli second latency and at most
> 2 milli second.
>

For read, write or both?


> That is not consistenly achievable with shared storage, to that end I
> added flash storage to a few hypervisors.
>

You could have flash on your shared storage.


> About 5% of my servers require this and are not that resource hungry to
> require a dedicated physical server.
> That same 5% also has no requirement to be migrated if a host fails.
>
> So in short I have 5 heavy hosts running ovirt with a shared storage
> domain on NFS for 95% of my servers.
> All running fine, but I am now unable to run my remaining 5%.
>

Perhaps, if there are no HA requirements, those VMs with local domain needs
can be in their own DC, a local one? If it's just 5%, shouldn't be much of
an effort?

>
> To finish up my summary I have been testing various virtualization
> technologies, like VmWare and Hyper-V.
> They allow such configurations as I mentioned.
>
> I already had some chat on irc with various guys and they suggested that I
> put this on the mailing list, so here goes.
>
> My suggestion would be to evoluate from SPM to SDM.
>

Easier said than done... We have worked on this for quite some time, it's
not as easy as one might think.


> SDM stands for Storage Domain Manager.
> This would create the possibility to have all nodes in the datacenter
> participate in the storage handling.
> A extra benefit would be that local storage could be added.
>
> What do you think?
>

There are other use cases we think flash on the host can be used, some may
be of use for your use case.
For example, dm-cache[1].

We are still looking at this. I think Gluster already can make use it for
cache, for example.
Y.

[1] https://people.redhat.com/mskinner/rhug/q1.2016/dm-cache.pdf


>
> Thanks for reading.
>
> Kind regards,
> Mike van Goor
> ___
> Users mailing list
> Users@ovirt.org
> http://lists.ovirt.org/mailman/listinfo/users
>
___
Users mailing list
Users@ovirt.org
http://lists.ovirt.org/mailman/listinfo/users

Re: [ovirt-users] Local and Shared storage in same datacenter

[Barak Korren]

> While I understand that this makes having local storage impossible, I
> believe there is a use case to have local storage in a shared storage
> datacenter.
> Consider the following:
> I have a few applications that require 1 milli second latency and at most 2
> milli second.
> That is not consistenly achievable with shared storage, to that end I added
> flash storage to a few hypervisors.
> About 5% of my servers require this and are not that resource hungry to
> require a dedicated physical server.
> That same 5% also has no requirement to be migrated if a host fails.
>

I mentioned this in another thread already, please look into the VDSM
scratchpad hook. It will allow you to attach (files on) the local SSDs
(as disks) to VMs when those VMs start up. It will perfectly meet your
use case if you only need to keep the data on the SSD while the VM is
up.


-- 
Barak Korren
bkor...@redhat.com
RHEV-CI Team
___
Users mailing list
Users@ovirt.org
http://lists.ovirt.org/mailman/listinfo/users

[ovirt-users] Local and Shared storage in same datacenter

[Mike (maillinglists)]

Hi guys,

There have been a few related questions already that I could find, but I 
did not find anything relating to my specific use case.


Currently it is not possible to mix local storage with shared storage in 
the same datacenter.

The reason seems to be because of the storage pool manager (SPM).
This is a role in the datacenter provided to one specific host.

While I understand that this makes having local storage impossible, I 
believe there is a use case to have local storage in a shared storage 
datacenter.

Consider the following:
I have a few applications that require 1 milli second latency and at 
most 2 milli second.
That is not consistenly achievable with shared storage, to that end I 
added flash storage to a few hypervisors.
About 5% of my servers require this and are not that resource hungry to 
require a dedicated physical server.

That same 5% also has no requirement to be migrated if a host fails.

So in short I have 5 heavy hosts running ovirt with a shared storage 
domain on NFS for 95% of my servers.

All running fine, but I am now unable to run my remaining 5%.

To finish up my summary I have been testing various virtualization 
technologies, like VmWare and Hyper-V.

They allow such configurations as I mentioned.

I already had some chat on irc with various guys and they suggested that 
I put this on the mailing list, so here goes.


My suggestion would be to evoluate from SPM to SDM.
SDM stands for Storage Domain Manager.
This would create the possibility to have all nodes in the datacenter 
participate in the storage handling.

A extra benefit would be that local storage could be added.

What do you think?

Thanks for reading.

Kind regards,
Mike van Goor
___
Users mailing list
Users@ovirt.org
http://lists.ovirt.org/mailman/listinfo/users

Re: [ovirt-users] Local storage & shared in same cluster

[Barak Korren]

On 27 October 2016 at 14:58, Sebastian Greco  wrote:
> Ohh thx for the answer :) That's why the running VMs on that particular
> storage would only run on that host. One can do that in other hypervisors
> like vsphere or hyper-v. It would be a nice thing to have in rhv too imho.
>

Well, you can partially do this, with the scratch-pad hook, if you
don't care about the local data going away when the VM shuts down.

https://github.com/oVirt/vdsm/tree/master/vdsm_hooks/scratchpad

We've been successfully using this in ovirt CI to keep intermediate
build process I/O out of our central storage.

It seems to me that short-lived non-migrating but stateful VMs are not
very interesting as a use case for RHV customers. When you run you
databases and applications on RHV, it seems to me you typically want
your data to stick around and your VMs to survive a hypervisor crash.
Then again, I do not have any data to substantiate or negate this
claim.


-- 
Barak Korren
bkor...@redhat.com
RHEV-CI Team
___
Users mailing list
Users@ovirt.org
http://lists.ovirt.org/mailman/listinfo/users

Re: [ovirt-users] oVirt 4.0 Reporting

[Yedidyah Bar David]

On Sat, Oct 29, 2016 at 12:09 PM,   wrote:
> Hello oVirt guru`s!
>
> In the first half of 2016 was the following information: 
> http://lists.ovirt.org/pipermail/users/2016-June/040549.html
>
> Are there any news about the built-in oVirt 4.0 reporting mechanism?

It's called now "Dashboard". It's the default first screen you see after
admin login in 4.0, you can't miss it.

> Is there any chance that we will see reports in the oVirt web console?

If you refer to the ovirt-engine-reports that was available in 3.6, then
the answer is No. If that's exactly what you need, you should stay in
3.6 for now. In principle you can setup a 3.6 ovirt-engine-reports machine
to work against a 4.0 engine+dwh, it might work, I am not aware of anyone
doing this.

Please check the following for 4.1+ plans in this area:

http://www.ovirt.org/develop/release-management/features/engine/metrics-store/

Best,
-- 
Didi
___
Users mailing list
Users@ovirt.org
http://lists.ovirt.org/mailman/listinfo/users