[Ocfs2-users] RES: How do I resize Ocfs2 filesystem using Pacemaker stack?

[Gang He]

Hello Carlos,


>>> 
> Hello Gang,
> Tank you for your fast answer.
> 
>>Hello Carlos,
>>>>> 
>>> Hi all.
>>> 
>>> We have a OCFS2 clustered file system running with Pacemaker stack 
>>> over a Dell storage.
>>> Now we need to enlarge the size of the file system. We already have 
>>> grown it at the storage and with fdisk it was recreated the partition 
>>> to allocate the new available size.
>>> 
>>> To resize the disk I umounted the filesystem and left the DLM running.
>>> 
>>> When I tried to resize the OCFS2 filesystem with tunefs.ocfs2, it 
>>> couldn't find the cluster stack.
>>> 
>>> apolo:~ # tunefs.ocfs2 -S /dev/mapper/storage1-part2
>>> tunefs.ocfs2: Unable to access cluster service while closing device 
>>> "/dev/mapper/storage1-part2"
>>Could you use "crm_mon" comand to identify the current cluster status is
> OK?
> 
> The file system is mounted at boot when the Pacemaker starts, as it is
> configured on it.
> 
> Last updated: Mon Dec  4 15:23:11 2017
> Last change: Sat Dec  2 02:03:59 2017
> Stack: corosync
> Current DC: diana (2) - partition with quorum
> Version: 1.1.12-1.1.12.git20140904.266d5c2
> 2 Nodes configured
> 13 Resources configured
> 
> 
> Online: [ apolo diana ]
> 
> stonith_sbd     (stonith:external/sbd): Started diana
>  Clone Set: base-clone [dlm]
>      Started: [ apolo diana ]
>  Clone Set: c-clusterfs [clusterfs]
>      Started: [ apolo diana ]
>  Clone Set: web_server [apache]
>      Started: [ apolo diana ]
>  Clone Set: c-ip_httpd [ip_httpd]
>      Started: [ apolo diana ]
> 
>>Next, could you mount this partition first? just want to know if this
> partition is integrated, since most people use LVM to extend the disk.  
> 
> This is how it is mounted
> apolo:~ # mount | grep ocfs2 
> /dev/mapper/storage1-part2 on /export type ocfs2
> (rw,noatime,heartbeat=none,nointr,data=ordered,errors=remount-ro,atime_quant
> um=60,localalloc=1,cluster_stack=pcmk,coherency=full,user_xattr,acl)
> 
>>Then do the file system extension command.
> 
> With the filesystem mounted the same issue happens.
> apolo:~ # tunefs.ocfs2 -S /dev/mapper/storage1-part2
> tunefs.ocfs2: Unable to access cluster service while closing device
> "/dev/mapper/storage1-part2"
> 
>>If still not work, please report a bug.
> 
> As openSUSE 13.2 isn't supported anymore, I don't know if they will take a
> look at this.
Base on your description, this should be considered as a bug, I will report in our SUSE bug system.
Then, we will investigate this issue.

> 
> Is there a way I can change the stack of the filesystem, extend it and
> change the stack back to Pacemaker?
o2cluster can help you do a little something, but I do NOT recommend you to change stack from pmck to o2cb/cman,
since there will be probably some problems on switching stack, this tool is used to tune some things.

Thanks
Gang

> 
> Regards,
> Carlos.