Storage edition
-
What do you mean by “exports file” ?
I think this is what i missed because I had already made the dev directory, copied the .mntcheck with cp and set to 777. -
/etc/exports
This is the file that tells what folders to open through nfs.
-
It’s working !
Yet, it still use “/images/dev”.
I think i missed something. -
In the init it uses /images/ and /images/dev. This is because from the client side it creates a folder on the client and mounts the NFS to that folder. Hopefully that makes sense.
-
And how make it use the right dev folder from the storage chosen to avoid transfer between drive ?
-
That’s done at tasking time. If the image is on the group/node that points at /images2/dev what’s sent to mount is IPOFFOG:/images2/dev
If it’s pointing on the group/node that’s looking at /images/dev it sends the mount point at: IPOFFOG:/images/dev
-
Doesn’t look to work, i’m uploading a computer with image set to “images2” and temp data are in the “/images/dev”
[ATTACH]1097[/ATTACH]
[url=“/_imported_xf_attachments/1/1097_fog.jpg?:”]fog.jpg[/url]
-
First i was think it was the created task as you said so i restarted it but it’s the same.
In syslog it’s requested to mount the /images2/dev but temp data end on /images/dev and image end in /images.
This is pretty crazy. -
Even deleting the default node to avoid to use it when images2 node is active, fog keeps trying to use it.
Jul 3 17:12:19 srv-fog in.tftpd[16789]: tftp: client does not accept options
Jul 3 17:12:33 srv-fog rpc.mountd[1760]: authenticated mount request from 10.121.20.101:743 for /images2/dev (/images2/dev)edit :
I find out that i had to put the main node in first in exports file and not second.
Since it’s the same datastore, it will use the first it an access. -
This post is deleted!
