[Gluster-users] Geo-replication setup problem

Aravinda avishwan at redhat.com
Wed Oct 7 05:13:32 UTC 2015


Cool. We will work on resolving Ubuntu packaging issues. Thanks for 
reporting.

Data sync always happens via Gluster Mount, Connecting Slave node is 
just for distributing the load. Each Master nodes randomly connect to 
available Slave nodes but sync happens via mount.(It is not data copy 
Brick to Brick)
Since Master has 4 nodes and Slave has 2 nodes. 2 Master nodes are 
connected to 1 Slave node.

regards
Aravinda

On 10/07/2015 05:43 AM, Wade Fitzpatrick wrote:
> So I ran the following on palace and madonna and it appears to be 
> syncing properly now
>
> root at palace:~# sed -i -e 
> 's:/usr/libexec/glusterfs/gsyncd:/usr/lib/x86_64-linux-gnu/glusterfs/gsyncd:' 
> .ssh/authorized_keys
>
> root at james:~/georepsetup# gluster volume geo-replication status
>
> MASTER NODE    MASTER VOL    MASTER BRICK                    SLAVE 
> USER    SLAVE SLAVE NODE    STATUS     CRAWL STATUS    LAST_SYNCED
> ------------------------------------------------------------------------------------------------------------------------------------------------------
> james          static /data/gluster1/static/brick1    root 
> ssh://palace::static    palace        Active     Hybrid Crawl    N/A
> cupid          static /data/gluster1/static/brick2    root 
> ssh://palace::static madonna       Passive    N/A             N/A
> hilton         static /data/gluster1/static/brick3    root 
> ssh://palace::static    palace        Passive    N/A             N/A
> present        static /data/gluster1/static/brick4    root 
> ssh://palace::static madonna       Passive    N/A             N/A
>
>
> I am still a little confused by the SLAVE column though - should 
> "palace" in that column really be an A record such as "gluster-remote" 
> that has 2 addresses (for both palace and madonna)?
>
> On 7/10/2015 8:10 am, Wade Fitzpatrick wrote:
>>
>>
>>
>>
>>
>> Thanks for the response, it looks like the problem is
>>
>>
>>
>> [2015-10-06 19:08:49.547874] E
>> [resource(/data/gluster1/static/brick1):222:errlog] Popen: command
>> "ssh -oPasswordAuthentication=no -oStrictHostKeyChecking=no -i
>> /var/lib/glusterd/geo-replication/secret.pem -oControlMaster=auto -S
>> /tmp/gsyncd-aux-ssh-RX3NRr/eb9c0906d6265193f403278fc1489309.sock
>> root at palace /nonexistent/gsyncd --session-owner
>> 3f9f810d-a988-4914-a5ca-5bd7b251a273 -N --listen --timeout 120
>> gluster://localhost:static" returned with 127, saying:
>>
>> [2015-10-06 19:08:49.547976] E
>> [resource(/data/gluster1/static/brick1):226:logerr] Popen: ssh>
>> This system is monitored and logged. Any unauthorized usage will be
>> prosecuted.
>>
>> [2015-10-06 19:08:49.548062] E
>> [resource(/data/gluster1/static/brick1):226:logerr] Popen: ssh>
>> bash: /usr/libexec/glusterfs/gsyncd: No such file or directory
>>
>>
>>
>>
>> All servers are Ubuntu 15.04 running glusterfs 3.7.4 built on Sep
>> 1 2015 12:08:58 but /usr/libexec/glusterfs does not exist so is
>> this a packaging error?
>>
>>
>>
>>
>> On 6/10/2015 6:11 pm, Aravinda wrote:
>>
>>
>>>
>>>
>>> Please share the tracebacks/errors from the logs(Master nodes)
>>> /var/log/glusterfs/geo-replication/static/*.log
>>>
>>> regards
>>> Aravinda
>>>
>>> On 10/06/2015 01:19 PM, Wade
>>> Fitzpatrick wrote:
>>>
>>>
>>>>
>>>>
>>>> I am trying to set up geo-replication of a striped-replicate
>>>> volume. I used https://github.com/aravindavk/georepsetup
>>>> to configure the replication.
>>>>
>>>>
>>>>
>>>> root at james:~# gluster volume info
>>>>
>>>>
>>>>
>>>> Volume Name: static
>>>>
>>>> Type: Striped-Replicate
>>>>
>>>> Volume ID: 3f9f810d-a988-4914-a5ca-5bd7b251a273
>>>>
>>>> Status: Started
>>>>
>>>> Number of Bricks: 1 x 2 x 2 = 4
>>>>
>>>> Transport-type: tcp
>>>>
>>>> Bricks:
>>>>
>>>> Brick1: james:/data/gluster1/static/brick1
>>>>
>>>> Brick2: cupid:/data/gluster1/static/brick2
>>>>
>>>> Brick3: hilton:/data/gluster1/static/brick3
>>>>
>>>> Brick4: present:/data/gluster1/static/brick4
>>>>
>>>> Options Reconfigured:
>>>>
>>>> changelog.changelog: on
>>>>
>>>> geo-replication.ignore-pid-check: on
>>>>
>>>> geo-replication.indexing: on
>>>>
>>>> performance.readdir-ahead: on
>>>>
>>>>
>>>>
>>>> root at james:~# gluster volume geo-replication status
>>>>
>>>>
>>>>
>>>> MASTER NODE    MASTER VOL    MASTER BRICK
>>>> SLAVE USER    SLAVE                   SLAVE NODE STATUS
>>>> CRAWL STATUS    LAST_SYNCED
>>>>
>>>> ------------------------------------------------------------------------------------------------------------------------------------------------------
>>>>
>>>> james          static        /data/gluster1/static/brick1
>>>> root ssh://palace::static
>>>>
>>>> N/A           Created    N/A             N/A
>>>>
>>>> cupid          static        /data/gluster1/static/brick2
>>>> root ssh://palace::static
>>>>
>>>> N/A           Created    N/A             N/A
>>>>
>>>> hilton         static        /data/gluster1/static/brick3
>>>> root ssh://palace::static
>>>>
>>>> N/A           Created    N/A             N/A
>>>>
>>>> present        static        /data/gluster1/static/brick4
>>>> root ssh://palace::static
>>>>
>>>> N/A           Created    N/A             N/A
>>>>
>>>>
>>>>
>>>> So of the 4 bricks, data is striped over brick1 and brick3, also
>>>> brick1=brick2 is a mirror and brick3=brick4 is a mirror.
>>>> Therefore I have no need to geo-replicate bricks 2 and 4.
>>>>
>>>>
>>>>
>>>> At the other site, palace and madonna form a stripe volume (no
>>>> replication):
>>>>
>>>>
>>>>
>>>> root at palace:~# gluster volume info
>>>>
>>>>
>>>>
>>>> Volume Name: static
>>>>
>>>> Type: Stripe
>>>>
>>>> Volume ID: 0e91c6f2-3499-4fc4-9630-9da8b7f57db5
>>>>
>>>> Status: Started
>>>>
>>>> Number of Bricks: 1 x 2 = 2
>>>>
>>>> Transport-type: tcp
>>>>
>>>> Bricks:
>>>>
>>>> Brick1: palace:/data/gluster1/static/brick1
>>>>
>>>> Brick2: madonna:/data/gluster1/static/brick2
>>>>
>>>> Options Reconfigured:
>>>>
>>>> performance.readdir-ahead: on
>>>>
>>>>
>>>>
>>>> However, when I try to start geo-replication, it fails as below.
>>>>
>>>>
>>>>
>>>> root at james:~# gluster volume geo-replication static 
>>>> ssh://palace::static start
>>>>
>>>> Starting geo-replication session between static & 
>>>> ssh://palace::static has been
>>>> successful
>>>>
>>>> root at james:~# gluster volume geo-replication status
>>>>
>>>>
>>>>
>>>> MASTER NODE    MASTER VOL    MASTER BRICK
>>>> SLAVE USER    SLAVE                   SLAVE NODE STATUS
>>>> CRAWL STATUS    LAST_SYNCED
>>>>
>>>> -----------------------------------------------------------------------------------------------------------------------------------------------------
>>>>
>>>> james          static        /data/gluster1/static/brick1
>>>> root ssh://palace::static
>>>>
>>>> N/A           Faulty    N/A             N/A
>>>>
>>>> cupid          static        /data/gluster1/static/brick2
>>>> root ssh://palace::static
>>>>
>>>> N/A           Faulty    N/A             N/A
>>>>
>>>> hilton         static        /data/gluster1/static/brick3
>>>> root ssh://palace::static
>>>>
>>>> N/A           Faulty    N/A             N/A
>>>>
>>>> present        static        /data/gluster1/static/brick4
>>>> root ssh://palace::static
>>>>
>>>> N/A           Faulty    N/A             N/A
>>>>
>>>>
>>>>
>>>>
>>>>
>>>> What should I do to set this up properly so that
>>>> james:/data/gluster1/static/brick1 gets replicated to
>>>> palace:/data/gluster1/static/brick1 ; and
>>>> hilton:/data/gluster1/static/brick3 gets replicated to
>>>> madonna:/data/gluster1/static/brick2 ???
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>> -- 
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>> _______________________________________________
>>>> 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

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.gluster.org/pipermail/gluster-users/attachments/20151007/344c328c/attachment.html>


More information about the Gluster-users mailing list