Re: Multiple Cache Pool with Single Storage Pool

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

 



If your goal is to localize reads and writes to the same node as a given VM (at least that's how I read your intent), creating a cache tier across your hypervisor hosts will not achieve that result since you can expect your data to be distributed across the pool as directed by the CRUSH map.

You are correct that by using a local (host) persistent cache, you have effectively removed the ability to safely live-migrate.

-- 

Jason Dillaman 


----- Original Message ----- 

> From: "Lazuardi Nasution" <mrxlazuardin@xxxxxxxxx>
> To: "Jason Dillaman" <dillaman@xxxxxxxxxx>
> Cc: ceph-users@xxxxxxxxxxxxxx
> Sent: Monday, November 9, 2015 12:47:18 PM
> Subject: Re:  Multiple Cache Pool with Single Storage Pool

> Hi Jason,
> What is the worst case if I made cache pool from local ASDa owner by all
> compute nodes? Is using block cache inside VM has compatibility issue with
> live migration?
> Best regards,
> On Nov 9, 2015 9:04 PM, "Jason Dillaman" < dillaman@xxxxxxxxxx > wrote:

> > This is currently not a possibility, but there is active research into
> > providing improved/persistent client-side caching for RBD use-cases. In the
> > meantime as an alternative, you can expose a portion of your SSD to each VM
> > that needs higher IOPS and apply dm-cache / bcache on the RBD and SSD
> > volume
> > internal to the VM.
> 

> > --
> 

> > Jason Dillaman
> 

> > ----- Original Message -----
> 

> > > From: "Lazuardi Nasution" < mrxlazuardin@xxxxxxxxx >
> 
> > > To: ceph-users@xxxxxxxxxxxxxx
> 
> > > Sent: Sunday, November 8, 2015 12:34:16 PM
> 
> > > Subject:  Multiple Cache Pool with Single Storage Pool
> 

> > > Hi,
> 

> > > I'm new with CEPH cache tiering. Is it possible to have multiple cache
> > > pool
> 
> > > with single storage pool backend? For example I have some compute nodes
> > > with
> 
> > > its own local SSDs. I want each compute node has its own cache by using
> > > its
> 
> > > own local SSDs. The target is to minimize load to storage pool backend.
> 

> > > If yes, what will be used for RBD client pool name, cache pool name or
> 
> > > storage pool name? How to make sure that RBD client will only use cache
> > > pool
> 
> > > constructed with its own local SSDs, not other nodes local SSDs?
> 

> > > Frankly, I'm looking for RBD client cache on SSDs, but it seem has not
> 
> > > available yet.
> 

> > > Best regards,
> 

> > > _______________________________________________
> 
> > > ceph-users mailing list
> 
> > > ceph-users@xxxxxxxxxxxxxx
> 
> > > http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com
> 
_______________________________________________
ceph-users mailing list
ceph-users@xxxxxxxxxxxxxx
http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com



[Index of Archives]     [Information on CEPH]     [Linux Filesystem Development]     [Ceph Development]     [Ceph Large]     [Linux USB Development]     [Video for Linux]     [Linux Audio Users]     [Yosemite News]     [Linux Kernel]     [Linux SCSI]     [xfs]


  Powered by Linux