Re: unbalance

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



MS
On 28 Feb 2015 13:27, "Jorick Astrego" <j.astrego@xxxxxxxxxxx> wrote:
>
>
> On 02/28/2015 08:36 AM, M S Vishwanath Bhat wrote:
>>
>>
>>
>> On 28 February 2015 at 02:17, Jorick Astrego <j.astrego@xxxxxxxxxxx> wrote:
>>>
>>> Hi,
>>>
>>> When I have a 4 or 6 node distributed replicated volume, is there an
>>> easy way to unbalance the data? The rebalancing is very nice, but I was
>>> thinking of a scenario where I would have take half the nodes offline. I
>>> could shift part or all of the data on one replicated volume and stuff
>>> it all on the other replicated volume.
>>
>>
>> I'm not sure what you meant by "unbalance the data" ?
>>
>> Do you want to replace the bricks (backend disks)? Or do you want to move the data from one gluster volume to another gluster volume?
>>
>> * If you want to replace the bricks, there is "gluster replace-brick". But there are some known issues with it. I would suggest "remove-brick" and then "add-brick" and rebalance instead.
>
> Say I have two racks with a distributed replicated setup
>
>
>
>
> And I need to take server3 and server4 offline and I don't have replacement bricks. How do I move File 2 to  Replicated Volume 0 while the file is in use (VM storage or big MP4)
If you do remove brick of replicate volume 1, it will move (rebalance) ask the data to replicate volume 0.
Note that it makes distribute replicate volume to pure replicate.

>
>
>> * If you want to move the data from one gluster volume to another, you can just rsync the data from mountpoint of volume1 to mountpoint of volume2.
>
> If I rsync to the other volume, will gluster dynamically remap the file and keep serving it? What about the old file left over on Server3 and Server4

Okay, when I say volume, I actually meant mountpoint of one Gluster volume to another Gluster volume mountpoint. But looks like that's not what you want. Sorry for the confusion.

Best Regards,
MS
>
>
>> HTH
>>
>> //MS
>>
>>
>>
>>>
>>> There would have to be enough space and performance will suffer a lot,
>>> but at least it will continue to be available.
>>>
>>>
>
>
>
>
> Met vriendelijke groet, With kind regards,
>
> Jorick Astrego
>
> Netbulae Virtualization Experts
> ________________________________
> Tel: 053 20 30 270
> info@xxxxxxxxxxx
> Staalsteden 4-3A
> KvK 08198180
> Fax: 053 20 30 271
> www.netbulae.eu
> 7547 TA Enschede
> BTW NL821234584B01
>
> ________________________________
>

_______________________________________________
Gluster-users mailing list
Gluster-users@xxxxxxxxxxx
http://www.gluster.org/mailman/listinfo/gluster-users

[Index of Archives]     [Gluster Development]     [Linux Filesytems Development]     [Linux ARM Kernel]     [Linux ARM]     [Linux Omap]     [Fedora ARM]     [IETF Annouce]     [Bugtraq]     [Linux OMAP]     [Linux MIPS]     [eCos]     [Asterisk Internet PBX]     [Linux API]

  Powered by Linux