Guys, Having a issue with NFS mount to a NetApp Export. Maybe someone had this issue and was able to get this working.
The permission on the Filer for the Export are, Any Host Read/Write, all host root access.
During the command: ./mount-nfs-store.pl --nfs <nfs_serverhostname>:/directory_to_use>
A error stop the process, the mount is accessible but data-va is not using it!
I can touch the nfs mount without a issue.
Any advice!
Thanks
Francisco
The mount script does the following:
1. Add the entry into /etc/fstab
2. (Start the nfs if not already)
3. Add the new volume to data-va
Did you see any errors while running this script?
Can you run this command on data-va node (as zimbra user, login as root and then "su - zimbra")
zmvolume -l (it should list a new store /opt/zimbra/store<int>)
And that folder should match the entry in the /etc/fstab
Thanks.
I can see the mount on the fstab
But dont show if you run zmvolume -l
Then I suggest running the command manually to add the zmvolume.
su - zimbra -c 'zmvolume -a -n store60 -t primaryMessage -p /opt/zimbra/store60 --compress false'
That should probably work.
So, run the command, Error occurred: directory does no exist or is not writable
But I can cp file to it and rm too. Any ideas!
Could this behavior be caused by a NFS/NetApp filer permissions issue? We get the same behavior here in the GSS lab with the same procedure where we do not have full access to the NFS store.
Here are my setting for that volume
Try it too adding the ip of the data-va
I open a ticket with NetApp (Ticket# 2004928757) to verify if there where any other setting for the NFS, and according to them is wide open. No restriction.
The just make sure that /opt/zimbra/store60 is readable/writable/executable by zimbra user.
Run the following commands:
chown zimbra:zimbra /opt/zimbra/store60
chmod ug+rwx /opt/zimbra/store60
And then restart the mailbox (/etc/init.d/zimbra restart)
That it! You are the man! After applying the chown and chmod command and running the
su - zimbra -c 'zmvolume -a -n store60 -t primaryMessage -p /opt/zimbra/store60 --compress false'
Volume 3 is created
zimbra@data-va:~> zmvolume -l
Volume id: 1
name: message1
type: primaryMessage
path: /opt/zimbra/store
compressed: false
current: true
Volume id: 2
name: index1
type: index
path: /opt/zimbra/index
compressed: false
current: true
Volume id: 3
name: store60
type: primaryMessage
path: /opt/zimbra/store60
compressed: false
current: false
Anything else I need to do? I see that current=False.
Another question, do I need to add this nfs mount to the additional data-va on a ha configuration?
You could set it to current using the zmvolume command:
My question is why was this neccessary? This is not documented anywhere in the public documentation, is this a know issue, a one off instance for this one customer, or somehting else?
Sorry i gave zimbra document. It is documented here:
But why was this necessary on a brand new deployment on 1.8?
If you want to use an external mount point, then we might have to do it.
i agree this must be better documented on the Administrator and HA. Neither Tech support knew about this! Regarding my last question, if I deploy more Data-va for HA, do I need to add this nfs mount to it? The documentation is not clear about this,
Could you give the location where that is referenced in the document as I did not come across it
For me, Does command must be already on the script that attach the NFS mount. That why the script fail in the beginning.
Now more things.
Supposedly I need to use the cli to tell the data-va to use the new NFS Mount, Documentation is missing exactly what I need to do and if I have more that one Data-va, do I need to add this nfs mount to the other ones?