Re: [Gluster-users] gluster remove-brick problem

[ghy...@qq.com]

Hi,
sorry for disturbing you again. I met another problem using GlusterFS.

I have several machines with CentOs 7.2 and RDMA hardware and software 
installed. And I can confirm RDMA works well on these machines.
I yum -y install glusterfs-rdma and glusterfs-server, and started 
glusterd . Created and started volume using transport type RDMA.

But I can not mount the glusterfs-server on another client machine. 
the log error is: W [MSGID: 103071] 
[rdma.c:1289:gf_rdma_cm_event_handler] 0-data-client-1: cma event 
RDMA_CM_EVENT_REJECTED, error 8

In the documentation of gluster, I saw "NOTE: As of now only FUSE 
client and gNFS server would support RDMA transport.”
so what is gNFS? How could I do to using RDMA transport type?
please give me some advice if possible 
Thanks you very much.

by starshine from beijing, China at 24 June, 2017.



> 在 2017年5月20日，00:48，Nithya Balachandran  写道：
> 
> Hi,
> 
> The rebalance could have failed because of any one of several reasons. You 
> would need to check the rebalance log for the volume to figure out why it 
> failed in this case. This should be /var/log/glusterfs/data-rebalance.log on 
> bigdata-dlp-server00.xg01.
> 
> I can take a look at the log if you send it across.
> 
> Regards,
> Nithya
> 
> On 19 May 2017 at 13:09, Pranith Kumar Karampuri  > wrote:
> Adding gluster-users, developers who work on distribute module of gluster.
> 
> On Fri, May 19, 2017 at 12:58 PM, 郭鸿岩(基础平台部)  > wrote:
> Hello,
> 
> I am a user of Gluster 3.8 from beijing, China.
> I met a problem.  I added a brick to a volume, but the brick is on 
> the / disk, the same disk with the linux OS.
> So I want to remove it. The volume is of distributed type without 
> replica.
> 
> I use commad: gluster volume remove-brick data 
> server1:/vdata/bricks/data start,
> It is ok. The command is successfully accepted by gluster. But then I 
> check status:  gluster volume remove-brick data server1:/vdata/bricks/data 
> status.
> IT showd:
> 
> Node Rebalanced-files size scanned failures skipped status run time 
> in h:m:s
>  - --- --- --- --- 
> ---  ———
>  bigdata-dlp-server00.xg01 0 0Bytes 0 1 0 failed 0:0:0
> 
> 0 files Scaned, 0 Bytes transfered and 1 failure.
> 
> Please give me some advice , How could I do to remove this brick without data 
> loss?
> 
> Thank you very much!
> 
> best wishes.
> 
> 
> by starshine.
> 
> 
> 
> -- 
> Pranith
> 

___
Gluster-users mailing list
Gluster-users@gluster.org
http://lists.gluster.org/mailman/listinfo/gluster-users

Re: [Gluster-users] gluster remove-brick problem

[基础平台部]

Thank you very much!
I have found the reason with the help of your advice.

I check the log carefully and thoroughly, and found  the cause of the failure 
is the server node can not connect the other peer ones because of DNS 
resolution.
After I configured the "IP hostnames” to /etc/hosts
It works as expect!

Thanks to everyone.  May you joy and happiness forever.

by starshine.




在 2017年5月20日，00:48，Nithya Balachandran 
mailto:nbala...@redhat.com>> 写道：

Hi,

The rebalance could have failed because of any one of several reasons. You 
would need to check the rebalance log for the volume to figure out why it 
failed in this case. This should be /var/log/glusterfs/data-rebalance.log on 
bigdata-dlp-server00.xg01.

I can take a look at the log if you send it across.

Regards,
Nithya

On 19 May 2017 at 13:09, Pranith Kumar Karampuri 
mailto:pkara...@redhat.com>> wrote:
Adding gluster-users, developers who work on distribute module of gluster.

On Fri, May 19, 2017 at 12:58 PM, 郭鸿岩(基础平台部) 
mailto:guohongyan...@didichuxing.com>> wrote:
Hello,

I am a user of Gluster 3.8 from beijing, China.
I met a problem.  I added a brick to a volume, but the brick is on the 
/ disk, the same disk with the linux OS.
So I want to remove it. The volume is of distributed type without 
replica.

I use commad: gluster volume remove-brick data 
server1:/vdata/bricks/data start,
It is ok. The command is successfully accepted by gluster. But then I 
check status:  gluster volume remove-brick data server1:/vdata/bricks/data 
status.
IT showd:

Node Rebalanced-files size scanned failures skipped status run time in 
h:m:s
 - --- --- --- --- --- 
 ———
 bigdata-dlp-server00.xg01 0 0Bytes 0 1 0 failed 0:0:0

0 files Scaned, 0 Bytes transfered and 1 failure.

Please give me some advice , How could I do to remove this brick without data 
loss?

Thank you very much!

best wishes.


by starshine.



--
Pranith


___
Gluster-users mailing list
Gluster-users@gluster.org
http://lists.gluster.org/mailman/listinfo/gluster-users

Re: [Gluster-users] gluster remove-brick problem

[Nithya Balachandran]

Hi,

The rebalance could have failed because of any one of several reasons. You
would need to check the rebalance log for the volume to figure out why it
failed in this case. This should be /var/log/glusterfs/data-rebalance.log
on bigdata-dlp-server00.xg01.

I can take a look at the log if you send it across.

Regards,
Nithya

On 19 May 2017 at 13:09, Pranith Kumar Karampuri 
wrote:

> Adding gluster-users, developers who work on distribute module of gluster.
>
> On Fri, May 19, 2017 at 12:58 PM, 郭鸿岩(基础平台部) <
> guohongyan...@didichuxing.com> wrote:
>
>> Hello,
>>
>> I am a user of Gluster 3.8 from beijing, China.
>> I met a problem.  I added a brick to a volume, but the brick is
>> on the / disk, the same disk with the linux OS.
>> So I want to remove it. The volume is of distributed type without
>> replica.
>>
>> I use commad: gluster volume remove-brick data
>> server1:/vdata/bricks/data start,
>> It is ok. The command is successfully accepted by gluster. But
>> then I check status:  gluster volume remove-brick data
>> server1:/vdata/bricks/data status.
>> IT showd:
>>
>> Node Rebalanced-files size scanned failures skipped status run
>> time in h:m:s
>>  - --- --- --- ---
>> ---  ———
>>  bigdata-dlp-server00.xg01 0 0Bytes 0 1 0 failed 0:0:0
>>
>> 0 files Scaned, 0 Bytes transfered and 1 failure.
>>
>> Please give me some advice , How could I do to remove this brick without
>> data loss?
>>
>> Thank you very much!
>>
>> best wishes.
>>
>>
>> by starshine.
>
>
>
>
> --
> Pranith
>
___
Gluster-users mailing list
Gluster-users@gluster.org
http://lists.gluster.org/mailman/listinfo/gluster-users

Re: [Gluster-users] gluster remove-brick problem

[Pranith Kumar Karampuri]

Adding gluster-users, developers who work on distribute module of gluster.

On Fri, May 19, 2017 at 12:58 PM, 郭鸿岩(基础平台部) 
wrote:

> Hello,
>
> I am a user of Gluster 3.8 from beijing, China.
> I met a problem.  I added a brick to a volume, but the brick is on
> the / disk, the same disk with the linux OS.
> So I want to remove it. The volume is of distributed type without
> replica.
>
> I use commad: gluster volume remove-brick data
> server1:/vdata/bricks/data start,
> It is ok. The command is successfully accepted by gluster. But
> then I check status:  gluster volume remove-brick data
> server1:/vdata/bricks/data status.
> IT showd:
>
> Node Rebalanced-files size scanned failures skipped status run
> time in h:m:s
>  - --- --- --- ---
> ---  ———
>  bigdata-dlp-server00.xg01 0 0Bytes 0 1 0 failed 0:0:0
>
> 0 files Scaned, 0 Bytes transfered and 1 failure.
>
> Please give me some advice , How could I do to remove this brick without
> data loss?
>
> Thank you very much!
>
> best wishes.
>
>
> by starshine.




-- 
Pranith
___
Gluster-users mailing list
Gluster-users@gluster.org
http://lists.gluster.org/mailman/listinfo/gluster-users