Login | Register For Free | Help
Search for: (Advanced)

Mailing List Archive: Linux-HA: Users

DRBD Concept Doubt

 

 

Linux-HA users RSS feed   Index | Next | Previous | View Threaded


netwarrior863 at gmail

May 20, 2012, 6:25 AM

Post #1 of 3 (253 views)
Permalink
DRBD Concept Doubt

Hi there list!

I've got a doubt regariding DRBD usage, at the moment I'm trying to
implement a HA systems with two nodes, is a easy and basic setup
Two servers, running oracle and LVM.

I configured once resource lest's say /dev/rootvg/myoracle-device on
both, this is working fine, I can perform a manual failover as follow

drdbadm secondary node1
umount /dev/drbd1

drdbadm primary node2
mount /dev/drbd1

This works fine and I have both serer sincronized, my problem or doubt
is, when the other node fails,lets say, I power it off, the node2 takes
primary
role, I do it manually, but I have the information Outdated and I loose
lots of information and I have to wait till node1 comes up to syncronize
with it.

So, does DRBD work like that? I thought DRBD was syncronizing in
backround to the other node to have both nodes with the same
information, PLEASE,
correct me if I'm wrong, cuz maybe this solution in not well implemented
configured or I missunderstood what'd DRBD is for.


Thanks for your time and support
Best regards




_______________________________________________
Linux-HA mailing list
Linux-HA [at] lists
http://lists.linux-ha.org/mailman/listinfo/linux-ha
See also: http://linux-ha.org/ReportingProblems


david at davidcoulson

May 20, 2012, 6:28 AM

Post #2 of 3 (244 views)
Permalink
Re: DRBD Concept Doubt [In reply to]

What is your clustering software and what is the configuration? Also
post your DRBD configuration and the output from cat /proc/drbd during
each stage of your testing which reproduces the issue.

Maybe post some kernel logs too would be helpful. Simply switching
pri/sec on DRBD won't cause a node to go outdated, unless you split
brain the environment.

David

On 5/20/12 9:25 AM, Net Warrior wrote:
> Hi there list!
>
> I've got a doubt regariding DRBD usage, at the moment I'm trying to
> implement a HA systems with two nodes, is a easy and basic setup
> Two servers, running oracle and LVM.
>
> I configured once resource lest's say /dev/rootvg/myoracle-device on
> both, this is working fine, I can perform a manual failover as follow
>
> drdbadm secondary node1
> umount /dev/drbd1
>
> drdbadm primary node2
> mount /dev/drbd1
>
> This works fine and I have both serer sincronized, my problem or doubt
> is, when the other node fails,lets say, I power it off, the node2 takes
> primary
> role, I do it manually, but I have the information Outdated and I loose
> lots of information and I have to wait till node1 comes up to syncronize
> with it.
>
> So, does DRBD work like that? I thought DRBD was syncronizing in
> backround to the other node to have both nodes with the same
> information, PLEASE,
> correct me if I'm wrong, cuz maybe this solution in not well implemented
> configured or I missunderstood what'd DRBD is for.
>
>
> Thanks for your time and support
> Best regards
>
>
>
>
> _______________________________________________
> Linux-HA mailing list
> Linux-HA [at] lists
> http://lists.linux-ha.org/mailman/listinfo/linux-ha
> See also: http://linux-ha.org/ReportingProblems
_______________________________________________
Linux-HA mailing list
Linux-HA [at] lists
http://lists.linux-ha.org/mailman/listinfo/linux-ha
See also: http://linux-ha.org/ReportingProblems


netwarrior863 at gmail

May 21, 2012, 5:34 PM

Post #3 of 3 (240 views)
Permalink
Re: DRBD Concept Doubt [In reply to]

El 05/20/2012 10:28 AM, David Coulson escribió:
> What is your clustering software and what is the configuration? Also
> post your DRBD configuration and the output from cat /proc/drbd during
> each stage of your testing which reproduces the issue.
>
> Maybe post some kernel logs too would be helpful. Simply switching
> pri/sec on DRBD won't cause a node to go outdated, unless you split
> brain the environment.
>
> David
>
> On 5/20/12 9:25 AM, Net Warrior wrote:
>> Hi there list!
>>
>> I've got a doubt regariding DRBD usage, at the moment I'm trying to
>> implement a HA systems with two nodes, is a easy and basic setup
>> Two servers, running oracle and LVM.
>>
>> I configured once resource lest's say /dev/rootvg/myoracle-device on
>> both, this is working fine, I can perform a manual failover as follow
>>
>> drdbadm secondary node1
>> umount /dev/drbd1
>>
>> drdbadm primary node2
>> mount /dev/drbd1
>>
>> This works fine and I have both serer sincronized, my problem or doubt
>> is, when the other node fails,lets say, I power it off, the node2 takes
>> primary
>> role, I do it manually, but I have the information Outdated and I loose
>> lots of information and I have to wait till node1 comes up to syncronize
>> with it.
>>
>> So, does DRBD work like that? I thought DRBD was syncronizing in
>> backround to the other node to have both nodes with the same
>> information, PLEASE,
>> correct me if I'm wrong, cuz maybe this solution in not well implemented
>> configured or I missunderstood what'd DRBD is for.
>>
>>
>> Thanks for your time and support
>> Best regards
>>
>>
>>
>>
>> _______________________________________________
>> Linux-HA mailing list
>> Linux-HA [at] lists
>> http://lists.linux-ha.org/mailman/listinfo/linux-ha
>> See also: http://linux-ha.org/ReportingProblems

Sorry, but there was a configuration issue, now I'm able to syncronize,
anyway I've have some question and I'd like to share my configuration
with you.


This is what I've got


resource myresource {

syncer {
rate 100M;
}

on node1 {
device /dev/drbd1;
disk /dev/rootvg/lv01;
address x.x.x.x:7789;
meta-disk /dev/rootvg/drbdmetadata[0];
}

on node2 {
device /dev/drbd1;
disk /dev/rootvg/lv01;
address x.x.x.x:7789;
meta-disk /dev/rootvg/drbdmetadata[0];
}

}

Cuz I did not want to play expanding the LV and loose the data I have on
it I just created another LV to store the metadada on it
/dev/rootvg/drbdmetadata
which is 512 MB in size.


I was able to perform a manual failover without data loss and I'm happy
with that, now I'd like to add another Logical Volume and now my 2
questions are

1 - Should I add the other LV to my config and restart or do I have to
make again a drbdadm create myresource ?
2 - Is drbdmetadata related to my LV size? any considerations on this
regarding it's size?




Thanks for your time and support
Best Regards



_______________________________________________
Linux-HA mailing list
Linux-HA [at] lists
http://lists.linux-ha.org/mailman/listinfo/linux-ha
See also: http://linux-ha.org/ReportingProblems

Linux-HA users RSS feed   Index | Next | Previous | View Threaded
 
 


Interested in having your list archived? Contact Gossamer Threads
 
  Web Applications & Managed Hosting Powered by Gossamer Threads Inc.