[Gluster-devel] Fwd: gluster-3.1.4 - testing of replication feature ??

[Pavan]

On Tuesday 10 May 2011 12:09 PM, shashidhar v wrote:
>
> Hi,
>
>
> I have configured glusterfs distributed replicated volume with two
> storage servers (eg: pvfs2 and pvfs3 )

First of all, your setup can be better. If you really want to use 
distributed-replicate, you need a minimum of 4 servers. And it is 
recommended to have the same space on each of the 4 servers (at least 
between replicate pairs if not all).

You have a distributed-replicate with only 2 servers - One with 6.4 GB 
of space and one with 4.5 GB. So, this results in a simple 2 node 
replicate setup where one server is a replica of the other. The maximum 
available space (max of 6.4 and 4.5) is what is being reported on the 
mount point.
Once you start copying files whose cumulative size exceeds 4.5 GB, the 
second server (which is a replica of the first) cannot accommodate any 
more of your files. That is why you are seeing an "out of space" error.

Pavan

>
> [root at pvfs2 ~]#  gluster volume create test-vol2 replica 2 transport tcp
> 172.x.x.x:/opt/glusterfs-pvfs2disk1 172.x.x.x:/opt/glusterfs-pvfs3disk1
> Creation of volume test-vol2 has been successful. Please start the
> volume to access data.
>
> [root at pvfs2 ~]# gluster volume start test-vol2
> Starting volume test-vol2 has been successful
>
>
> storage server1:/opt/glusterfs-pvfs2disk1 has around 6.4GB and
> storage2:/opt/glusterfs-pvfs3disk1 has around 4.5GB of space
>
> And a client node (pvfs4) NFS mounted to the server1(pvfs2)
> [root at pvfs4 glusterfs]# mount -t nfs 172.x.x.x:/test-vol2 /mnt/glusterfs
>
> And the mounted path shows the total space as 6.4GB i.e.,server1
> capacity only ( second server storage is not included )
>
> How can I test the replication feature using these two servers , KIndly
> help me to do the testing of glusterfs with sample examples or
> benchmarks using tranport protocols as tcp and rdma . ( I am able to
> configure the storage volumes using tranport as tcp or rdma )
>
> I tried "cp command"  to copy the files around total space of 6.4GB to
> the nfs mounted directory in client machine(pvfs4) , after that it shows
> the error as no space on device, I have checked that all the copied
> files are physically stored under server1(pvfs2) only but not on second
> server (pvfs3)
>
> And there are some logs on second server(pvfs3)
> [root at pvfs3 ~]# vi /var/log/glusterfs/bricks/opt-glusterfs-pvfs3disk1.log
> [2011-05-06 12:52:44.179880] E [common-utils.c:124:gf_resolve_ip6]
> 0-resolver: getaddrinfo failed (Name or service not known)
> [2011-05-06 12:52:44.179934] E
> [name.c:251:af_inet_client_get_remote_sockaddr] 0-glusterfs: DNS
> resolution failed on host localhost
> [2011-05-06 12:52:44.179984] E [glusterfsd-mgmt.c:452:mgmt_rpc_notify]
> 0-glusterfsd-mgmt: failed to connect with remote-host: Success
> [2011-05-06 12:52:44.180309] E [common-utils.c:124:gf_resolve_ip6]
> 0-resolver: getaddrinfo failed (Name or service not known)
> [2011-05-06 12:52:44.180326] E
> [name.c:251:af_inet_client_get_remote_sockaddr] 0-glusterfs: DNS
> resolution failed on host localhost
> [2011-05-06 12:52:44.180364] I [glusterfsd.c:700:cleanup_and_exit]
> 0-glusterfsd: shutting down
>
> Kindly help me to test the glusterfs with examples or becnhmakrs for
> both tcp and rdma
>
>
> Thanks & Regards,
> shashi
>
>
>
>
>
>
>
>
>
>
> _______________________________________________
> Gluster-devel mailing list
> Gluster-devel at nongnu.org
> https://lists.nongnu.org/mailman/listinfo/gluster-devel