[Gluster-devel] Gluster Sharding and Geo-replication

Krutika Dhananjay kdhananj at redhat.com
Thu Sep 3 10:42:01 UTC 2015


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

> From: "Sahina Bose" <sabose at redhat.com>
> To: "Krutika Dhananjay" <kdhananj at redhat.com>, "Shyam" <srangana at redhat.com>
> Cc: "Gluster Devel" <gluster-devel at gluster.org>
> Sent: Thursday, September 3, 2015 3:56:10 PM
> Subject: Re: [Gluster-devel] Gluster Sharding and Geo-replication

> On 09/03/2015 12:13 PM, Krutika Dhananjay wrote:

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

> > > From: "Shyam" <srangana at redhat.com>
> > 
> 
> > > To: "Krutika Dhananjay" <kdhananj at redhat.com>
> > 
> 
> > > Cc: "Aravinda" <avishwan at redhat.com> , "Gluster Devel"
> > > <gluster-devel at gluster.org>
> > 
> 
> > > Sent: Wednesday, September 2, 2015 11:13:55 PM
> > 
> 
> > > Subject: Re: [Gluster-devel] Gluster Sharding and Geo-replication
> > 
> 

> > > On 09/02/2015 10:47 AM, Krutika Dhananjay wrote:
> > 
> 
> > > >
> > 
> 
> > > >
> > 
> 
> > > > ------------------------------------------------------------------------
> > 
> 
> > > >
> > 
> 
> > > > *From: *"Shyam" <srangana at redhat.com>
> > 
> 
> > > > *To: *"Aravinda" <avishwan at redhat.com> , "Gluster Devel"
> > 
> 
> > > > <gluster-devel at gluster.org>
> > 
> 
> > > > *Sent: *Wednesday, September 2, 2015 8:09:55 PM
> > 
> 
> > > > *Subject: *Re: [Gluster-devel] Gluster Sharding and Geo-replication
> > 
> 
> > > >
> > 
> 
> > > > On 09/02/2015 03:12 AM, Aravinda wrote:
> > 
> 
> > > > > Geo-replication and Sharding Team today discussed about the approach
> > 
> 
> > > > > to make Sharding aware Geo-replication. Details are as below
> > 
> 
> > > > >
> > 
> 
> > > > > Participants: Aravinda, Kotresh, Krutika, Rahul Hinduja, Vijay Bellur
> > 
> 
> > > > >
> > 
> 
> > > > > - Both Master and Slave Volumes should be Sharded Volumes with same
> > 
> 
> > > > > configurations.
> > 
> 
> > > >
> > 
> 
> > > > If I am not mistaken, geo-rep supports replicating to a non-gluster
> > 
> 
> > > > local FS at the slave end. Is this correct? If so, would this
> > 
> 
> > > > limitation
> > 
> 
> > > > not make that problematic?
> > 
> 
> > > >
> > 
> 
> > > > When you state *same configuration*, I assume you mean the sharding
> > 
> 
> > > > configuration, not the volume graph, right?
> > 
> 
> > > >
> > 
> 
> > > > That is correct. The only requirement is for the slave to have shard
> > 
> 
> > > > translator (for, someone needs to present aggregated view of the file
> > > > to
> > 
> 
> > > > the READers on the slave).
> > 
> 
> > > > Also the shard-block-size needs to be kept same between master and
> > 
> 
> > > > slave. Rest of the configuration (like the number of subvols of
> > > > DHT/AFR)
> > 
> 
> > > > can vary across master and slave.
> > 
> 

> > > Do we need to have the sharded block size the same? As I assume the file
> > 
> 
> > > carries an xattr that contains the size it is sharded with
> > 
> 
> > > (trusted.glusterfs.shard.block-size), so if this is synced across, it
> > 
> 
> > > would do. If this is true, what it would mean is that "a sharded volume
> > 
> 
> > > needs a shard supported slave to ge-rep to".
> > 
> 

> > Yep. Even I feel it should probably not be necessary to enforce
> > same-shard-size-everywhere as long as shard translator on the slave takes
> > care not to further "shard" the individual shards gsyncD would write to, on
> > the slave volume.
> 
> > This is especially true if different files/images/vdisks on the master
> > volume
> > are associated with different block sizes.
> 
> > This logic has to be built into the shard translator based on parameters
> > (client-pid, parent directory of the file being written to).
> 
> > What this means is that shard-block-size attribute on the slave would
> > essentially be a don't-care parameter. I need to give all this some more
> > thought though.
> 

> I think this may help with coping with changes in shard block size
> configuration in master. Otherwise, once user changes the shard block size
> in master, the slave will be affected.
> Are there any other shard volume options that if changed on master, would
> affect slave? How do we ensure master and slave are in sync w.r.t the shard
> configuration?
The shard-block-size itself will be tied to a file since its creation in the form of an xattr. Subsequent changes to the block size of a volume should have no impact on already existing sharded files. 
So it is probably not necessary for the shard size option to even be synced to the slave volume, as long as the block-size xattr is correctly synced to the slave's copy of the file. 

Also, shard-block-size is the only tunable option as of today in shard translator. In future, there could be few more options introduced for 
performance optimizations but that should probably not mean anything to the slave volume. 

-Krutika 

> > -Krutika
> 

> > > >
> > 
> 
> > > > -Krutika
> > 
> 
> > > >
> > 
> 
> > > >
> > 
> 
> > > >
> > 
> 
> > > > > - In Changelog record changes related to Sharded files also. Just
> > 
> 
> > > > like
> > 
> 
> > > > > any regular files.
> > 
> 
> > > > > - Sharding should allow Geo-rep to list/read/write Sharding internal
> > 
> 
> > > > > Xattrs if Client PID is gsyncd(-1)
> > 
> 
> > > > > - Sharding should allow read/write of Sharded files(that is in
> > 
> 
> > > > .shards
> > 
> 
> > > > > directory) if Client PID is GSYNCD
> > 
> 
> > > > > - Sharding should return actual file instead of returning the
> > 
> 
> > > > > aggregated content when the Main file is requested(Client PID
> > 
> 
> > > > > GSYNCD)
> > 
> 
> > > > >
> > 
> 
> > > > > For example, a file f1 is created with GFID G1.
> > 
> 
> > > > >
> > 
> 
> > > > > When the file grows it gets sharded into chunks(say 5 chunks).
> > 
> 
> > > > >
> > 
> 
> > > > > f1 G1
> > 
> 
> > > > > .shards/G1.1 G2
> > 
> 
> > > > > .shards/G1.2 G3
> > 
> 
> > > > > .shards/G1.3 G4
> > 
> 
> > > > > .shards/G1.4 G5
> > 
> 
> > > > >
> > 
> 
> > > > > In Changelog, this is recorded as 5 different files as below
> > 
> 
> > > > >
> > 
> 
> > > > > CREATE G1 f1
> > 
> 
> > > > > DATA G1
> > 
> 
> > > > > META G1
> > 
> 
> > > > > CREATE G2 PGS/G1.1
> > 
> 
> > > > > DATA G2
> > 
> 
> > > > > META G1
> > 
> 
> > > > > CREATE G3 PGS/G1.2
> > 
> 
> > > > > DATA G3
> > 
> 
> > > > > META G1
> > 
> 
> > > > > CREATE G4 PGS/G1.3
> > 
> 
> > > > > DATA G4
> > 
> 
> > > > > META G1
> > 
> 
> > > > > CREATE G5 PGS/G1.4
> > 
> 
> > > > > DATA G5
> > 
> 
> > > > > META G1
> > 
> 
> > > > >
> > 
> 
> > > > > Where PGS is GFID of .shards directory.
> > 
> 
> > > > >
> > 
> 
> > > > > Geo-rep will create these files independently in Slave Volume and
> > 
> 
> > > > > syncs Xattrs of G1. Data can be read only when all the chunks are
> > 
> 
> > > > > synced to Slave Volume. Data can be read partially if main/first file
> > 
> 
> > > > > and some of the chunks synced to Slave.
> > 
> 
> > > > >
> > 
> 
> > > > > Please add if I missed anything. C & S Welcome.
> > 
> 
> > > > >
> > 
> 
> > > > > regards
> > 
> 
> > > > > Aravinda
> > 
> 
> > > > >
> > 
> 
> > > > > On 08/11/2015 04:36 PM, Aravinda wrote:
> > 
> 
> > > > >> Hi,
> > 
> 
> > > > >>
> > 
> 
> > > > >> We are thinking different approaches to add support in
> > 
> 
> > > > Geo-replication
> > 
> 
> > > > >> for Sharded Gluster Volumes[1]
> > 
> 
> > > > >>
> > 
> 
> > > > >> *Approach 1: Geo-rep: Sync Full file*
> > 
> 
> > > > >> - In Changelog only record main file details in the same brick
> > 
> 
> > > > >> where it is created
> > 
> 
> > > > >> - Record as DATA in Changelog whenever any addition/changes
> > 
> 
> > > > to the
> > 
> 
> > > > >> sharded file
> > 
> 
> > > > >> - Geo-rep rsync will do checksum as a full file from mount and
> > 
> 
> > > > >> syncs as new file
> > 
> 
> > > > >> - Slave side sharding is managed by Slave Volume
> > 
> 
> > > > >> *Approach 2: Geo-rep: Sync sharded file separately*
> > 
> 
> > > > >> - Geo-rep rsync will do checksum for sharded files only
> > 
> 
> > > > >> - Geo-rep syncs each sharded files independently as new files
> > 
> 
> > > > >> - [UNKNOWN] Sync internal xattrs(file size and block count)
> > 
> 
> > > > in the
> > 
> 
> > > > >> main sharded file to Slave Volume to maintain the same state as
> > 
> 
> > > > in Master.
> > 
> 
> > > > >> - Sharding translator to allow file creation under .shards
> > 
> 
> > > > dir for
> > 
> 
> > > > >> gsyncd. that is Parent GFID is .shards directory
> > 
> 
> > > > >> - If sharded files are modified during Geo-rep run may end up
> > 
> 
> > > > stale
> > 
> 
> > > > >> data in Slave.
> > 
> 
> > > > >> - Files on Slave Volume may not be readable unless all sharded
> > 
> 
> > > > >> files sync to Slave(Each bricks in Master independently sync
> > 
> 
> > > > files to
> > 
> 
> > > > >> slave)
> > 
> 
> > > > >>
> > 
> 
> > > > >> First approach looks more clean, but we have to analize the Rsync
> > 
> 
> > > > >> checksum performance on big files(Sharded in backend, accessed
> > 
> 
> > > > as one
> > 
> 
> > > > >> big file from rsync)
> > 
> 
> > > > >>
> > 
> 
> > > > >> Let us know your thoughts. Thanks
> > 
> 
> > > > >>
> > 
> 
> > > > >> Ref:
> > 
> 
> > > > >> [1]
> > 
> 
> > > > >>
> > 
> 
> > > > http://www.gluster.org/community/documentation/index.php/Features/sharding-xlator
> > 
> 
> > > > >> --
> > 
> 
> > > > >> regards
> > 
> 
> > > > >> Aravinda
> > 
> 
> > > > >>
> > 
> 
> > > > >>
> > 
> 
> > > > >> _______________________________________________
> > 
> 
> > > > >> Gluster-devel mailing list
> > 
> 
> > > > >> Gluster-devel at gluster.org
> > 
> 
> > > > >> http://www.gluster.org/mailman/listinfo/gluster-devel
> > 
> 
> > > > >
> > 
> 
> > > > >
> > 
> 
> > > > >
> > 
> 
> > > > > _______________________________________________
> > 
> 
> > > > > Gluster-devel mailing list
> > 
> 
> > > > > Gluster-devel at gluster.org
> > 
> 
> > > > > http://www.gluster.org/mailman/listinfo/gluster-devel
> > 
> 
> > > > >
> > 
> 
> > > > _______________________________________________
> > 
> 
> > > > Gluster-devel mailing list
> > 
> 
> > > > Gluster-devel at gluster.org
> > 
> 
> > > > http://www.gluster.org/mailman/listinfo/gluster-devel
> > 
> 
> > > >
> > 
> 
> > > >
> > 
> 

> > _______________________________________________
> 
> > Gluster-devel mailing list Gluster-devel at gluster.org
> > http://www.gluster.org/mailman/listinfo/gluster-devel
> 
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.gluster.org/pipermail/gluster-devel/attachments/20150903/97ab9c07/attachment-0001.html>


More information about the Gluster-devel mailing list