[Gluster-users] Setting up geo replication with GlusterFS 3.6.5
Saravanakumar Arumugam
sarumuga at redhat.com
Wed Sep 16 06:42:40 UTC 2015
Hi,
Replies inline.
On 09/16/2015 01:34 AM, ML mail wrote:
> Thanks for your detailed example. Based on that it looks like my issue is SSH based. Now I have the following two SSH related questions:
>
> 1) The setup of a SSH passwordless account on the slave, does it need to be using the same SSH public key as stored by GlusterFS in the
> /var/lib/glusterd/geo-replication directory? or can I simply generate my own with ssh-keygen?
You can simply generate on your own with ssh-keygen.
You should be able to login from Master node to Slave node without password.
(<master># ssh root@<slave_ip> ) thats it.
>
> 2) is it possible to use another user than root for geo-replication with GlusterFS v3.6?
It is supported in 3.6.5. It involves more steps in addition to below
mentioned(which is for root user).
Please refer the link which you have
mentioned.(http://www.gluster.org/pipermail/gluster-users.old/2015-January/020080.html)
>
> Regards
> ML
>
>
>
> On Tuesday, September 15, 2015 9:16 AM, Saravanakumar Arumugam <sarumuga at redhat.com> wrote:
> Hi,
> You are right, This tool may not be compatible with 3.6.5.
>
> I have tried myself with 3.6.5, but faced this error.
> ==========================
> georepsetup tv1 gfvm3 tv2
> Geo-replication session will be established between tv1 and gfvm3::tv2
> Root password of gfvm3 is required to complete the setup. NOTE: Password
> will not be stored.
>
> root at gfvm3's password:
> [ OK] gfvm3 is Reachable(Port 22)
> [ OK] SSH Connection established root at gfvm3
> [ OK] Master Volume and Slave Volume are compatible (Version: 3.6.5)
> [ OK] Common secret pub file present at
> /var/lib/glusterd/geo-replication/common_secret.pem.pub
> [ OK] common_secret.pem.pub file copied to gfvm3
> [ OK] Master SSH Keys copied to all Up Slave nodes
> [ OK] Updated Master SSH Keys to all Up Slave nodes authorized_keys file
> [NOT OK] Failed to Establish Geo-replication Session
> Command type not found while handling geo-replication options
> [root at gfvm3 georepsetup]#
> ==========================
> So, some more changes required in this tool.
>
>
> Coming back to your question:
>
> I have setup geo-replication using the commands in 3.6.5.
> Please recheck all the commands (with necessary changes at your end).
>
> ====================================================================
> [root at gfvm3 georepsetup]#
> [root at gfvm3 georepsetup]# cat /etc/redhat-release
> Fedora release 21 (Twenty One)
> [root at gfvm3 georepsetup]#
> [root at gfvm3 georepsetup]# rpm -qa | grep glusterfs
> glusterfs-devel-3.6.5-1.fc21.x86_64
> glusterfs-3.6.5-1.fc21.x86_64
> glusterfs-rdma-3.6.5-1.fc21.x86_64
> glusterfs-fuse-3.6.5-1.fc21.x86_64
> glusterfs-server-3.6.5-1.fc21.x86_64
> glusterfs-debuginfo-3.6.5-1.fc21.x86_64
> glusterfs-libs-3.6.5-1.fc21.x86_64
> glusterfs-extra-xlators-3.6.5-1.fc21.x86_64
> glusterfs-geo-replication-3.6.5-1.fc21.x86_64
> glusterfs-api-3.6.5-1.fc21.x86_64
> glusterfs-api-devel-3.6.5-1.fc21.x86_64
> glusterfs-cli-3.6.5-1.fc21.x86_64
> [root at gfvm3 georepsetup]#
> [root at gfvm3 georepsetup]#
> [root at gfvm3 georepsetup]# service glusterd start
> Redirecting to /bin/systemctl start glusterd.service
> [root at gfvm3 georepsetup]#
> [root at gfvm3 georepsetup]#
> [root at gfvm3 georepsetup]# service glusterd status
> Redirecting to /bin/systemctl status glusterd.service
> ● glusterd.service - GlusterFS, a clustered file-system server
> Loaded: loaded (/usr/lib/systemd/system/glusterd.service; disabled)
> Active: active (running) since Tue 2015-09-15 12:19:32 IST; 4s ago
> Process: 2778 ExecStart=/usr/sbin/glusterd -p /var/run/glusterd.pid
> (code=exited, status=0/SUCCESS)
> Main PID: 2779 (glusterd)
> CGroup: /system.slice/glusterd.service
> └─2779 /usr/sbin/glusterd -p /var/run/glusterd.pid
> [root at gfvm3 georepsetup]#
> [root at gfvm3 georepsetup]#
> [root at gfvm3 georepsetup]# ps aux | grep glus
> root 2779 0.0 0.4 448208 17288 ? Ssl 12:19 0:00
> /usr/sbin/glusterd -p /var/run/glusterd.pid
> [root at gfvm3 georepsetup]#
> [root at gfvm3 georepsetup]#
> [root at gfvm3 georepsetup]# gluster volume create tv1
> gfvm3:/opt/volume_test/tv_1/b1 gfvm3:/opt/volume_test/tv_1/b2 force
> volume create: tv1: success: please start the volume to access data
> [root at gfvm3 georepsetup]#
> [root at gfvm3 georepsetup]#
> [root at gfvm3 georepsetup]# gluster volume create tv2
> gfvm3:/opt/volume_test/tv_2/b1 gfvm3:/opt/volume_test/tv_2/b2 force
> volume create: tv2: success: please start the volume to access data
> [root at gfvm3 georepsetup]#
> [root at gfvm3 georepsetup]#
> [root at gfvm3 georepsetup]# gluster volume start tv1
> volume start: tv1: success
> [root at gfvm3 georepsetup]#
> [root at gfvm3 georepsetup]#
> [root at gfvm3 georepsetup]# gluster volume start tv2
> volume start: tv2: success
> [root at gfvm3 georepsetup]#
> [root at gfvm3 georepsetup]#
> [root at gfvm3 georepsetup]# mount -t glusterfs gfvm3:/tv1 /mnt/master/
> [root at gfvm3 georepsetup]# mount -t glusterfs gfvm3:/tv2 /mnt/slave/
> [root at gfvm3 georepsetup]#
> [root at gfvm3 georepsetup]#
> [root at gfvm3 georepsetup]# gluster system:: execute gsec_create
> Common secret pub file present at
> /var/lib/glusterd/geo-replication/common_secret.pem.pub
> [root at gfvm3 georepsetup]#
> [root at gfvm3 georepsetup]#
> [root at gfvm3 georepsetup]# gluster volume geo-replication tv1 gfvm3::tv2
> create push-pem
> Creating geo-replication session between tv1 & gfvm3::tv2 has been
> successful
> [root at gfvm3 georepsetup]#
> [root at gfvm3 georepsetup]#
> [root at gfvm3 georepsetup]# gluster volume geo-replication tv1 gfvm3::tv2
> start
> Starting geo-replication session between tv1 & gfvm3::tv2 has been
> successful
> [root at gfvm3 georepsetup]#
> [root at gfvm3 georepsetup]#
> [root at gfvm3 georepsetup]# gluster volume geo-replication tv1 gfvm3::tv2
> status
>
> MASTER NODE MASTER VOL MASTER BRICK SLAVE
> STATUS CHECKPOINT STATUS CRAWL STATUS
> ----------------------------------------------------------------------------------------------------------------------------
> gfvm3 tv1 /opt/volume_test/tv_1/b1 gfvm3::tv2
> Initializing... N/A N/A
> gfvm3 tv1 /opt/volume_test/tv_1/b2 gfvm3::tv2
> Initializing... N/A N/A
> [root at gfvm3 georepsetup]#
> [root at gfvm3 georepsetup]#
> [root at gfvm3 georepsetup]#
> [root at gfvm3 georepsetup]#
> [root at gfvm3 georepsetup]# gluster volume geo-replication tv1 gfvm3::tv2
> status
>
> MASTER NODE MASTER VOL MASTER BRICK SLAVE STATUS
> CHECKPOINT STATUS CRAWL STATUS
> ----------------------------------------------------------------------------------------------------------------------
> gfvm3 tv1 /opt/volume_test/tv_1/b1 gfvm3::tv2
> Active N/A Changelog Crawl
> gfvm3 tv1 /opt/volume_test/tv_1/b2 gfvm3::tv2
> Active N/A Changelog Crawl
> [root at gfvm3 georepsetup]#
> [root at gfvm3 georepsetup]# cp /etc/hosts
> hosts hosts.allow hosts.deny
> [root at gfvm3 georepsetup]# cp /etc/hosts* /mnt/master; sleep 20; ls
> /mnt/slave/
> hosts hosts.allow hosts.deny
> [root at gfvm3 georepsetup]# ls /mnt/master
> hosts hosts.allow hosts.deny
> [root at gfvm3 georepsetup]# ls /mnt/slave/
> hosts hosts.allow hosts.deny
> [root at gfvm3 georepsetup]#
> ====================================================================
> One important step which I have NOT mentioned is you need to setup
> passwordless ssh.
>
> You need to use ssh-keygen and ssh-copy-id for having passwordless setup
> from one node in Master to one node in Slave.
> This needs to be carried out before this step: " gluster system::
> execute gsec_create"
> This needs to be done at the same NODE(at Master) where you execute
> geo-rep create command.
>
> You can find geo-replication related logs here
> :/var/log/glusterfs/geo-replication/
> Please share the logs if you still face any issues.
>
> Thanks,
> Saravana
>
>
> On 09/14/2015 11:23 PM, ML mail wrote:
>> Yes I can ping the slave node with it's name and IP address, I've even entered manually its name in /etc/hosts.
>>
>> Does this nice python script also work for Gluster 3.6? The blog post only speaks about 3.7...
>>
>> Regards
>> ML
>>
>>
>>
>>
>> On Monday, September 14, 2015 9:38 AM, Saravanakumar Arumugam <sarumuga at redhat.com> wrote:
>> Hi,
>>
>> << Unable to fetch slave volume details. Please check the slave cluster
>> and slave volume. geo-replication command failed
>> Have you checked whether you are able to reach the Slave node from the
>> master node?
>>
>> There is a super simple way of setting up geo-rep written by Aravinda.
>> Refer:
>> http://blog.gluster.org/2015/09/introducing-georepsetup-gluster-geo-replication-setup-tool-2/
>>
>> Refer the README for both usual (root user based) and
>> mountbroker(non-root) setup details here:
>> https://github.com/aravindavk/georepsetup/blob/master/README.md
>>
>> Thanks,
>> Saravana
>>
>>
>>
>> On 09/13/2015 09:46 PM, ML mail wrote:
>>> Hello,
>>>
>>> I am using the following documentation in order to setup geo replication between two sites
>>> http://www.gluster.org/pipermail/gluster-users.old/2015-January/020080.html
>>>
>>> Unfortunately the step:
>>>
>>> gluster volume geo-replication myvolume gfsgeo at gfs1geo.domain.com::myvolume create push-pem
>>>
>>> Fails with the following error:
>>>
>>> Unable to fetch slave volume details. Please check the slave cluster and slave volume.
>>> geo-replication command failed
>>>
>>> Any ideas?
>>>
>>> btw: the documentation
>>> http://gluster.readthedocs.org/en/latest/Administrator%20Guide/Geo%20Replication/index.html
>>> does not seem to work with GlusterFS 3.6.5 that's why I am using the other mentioned documentation. It fails at the mountbroker step (
>>> gluster system:: execute mountbroker opt mountbroker-root /var/mountbroker-root)
>>>
>>> Regards
>>> ML
>>> _______________________________________________
>>> Gluster-users mailing list
>>> Gluster-users at gluster.org
>>> http://www.gluster.org/mailman/listinfo/gluster-users
>> _______________________________________________
>> Gluster-users mailing list
>> Gluster-users at gluster.org
>> http://www.gluster.org/mailman/listinfo/gluster-users
> _______________________________________________
> Gluster-users mailing list
> Gluster-users at gluster.org
> http://www.gluster.org/mailman/listinfo/gluster-users
More information about the Gluster-users
mailing list