On Fri, 2 Feb 2007 20:18:11 -0600
"Christopher D. Clausen" <[EMAIL PROTECTED]> wrote:

> This may have been discussed before, but is anyone strongly against just 
> waiting for etch (the next Debian stable release) to be out before 
> migrating users?  This should provide newer packages for nearly 
> everything and may save in the long run, as trying to dist-upgrade with 
> users on the system won't be pretty and I don't think we want to keep 
> running sarge for too much longer after etch is out.

Yeah, I don't like this. 1st problem is time/money @ the colo facility,
and the second thing is, upgrading to etch isn't that much of a big deal
really.

Just 2 days ago I upgraded a long-running sarge system to etch with
no problems at all.

> I believe I talked with docelic on IRC about creating a seperate volume 
> for each user's db.  I'd suggest setting the initial quota to 50MB 
> (thats about 50000 K, which is what the AFS commands use.)  There was 
> also aa suggestion to create and add a system:backup group with rl ACLs 
> to every volume, in case some sort of file-based backup system is to be 
> used in the future.

Yes, we've added system:backup.

> 
> Additionally, I'd suggest granting the user at least rl access to this 
> volume.  They should be able to check the quota on the volume and it is 
> usally easier to simply cd into it and run fs lq than to remember to use 
> the vos listvol or vos examine command.

Right, please add/modify what I wrote on
http://wiki.hcoop.net/wiki/DaemonAdmin/MySQL
to have it work this way.

_______________________________________________
HCoop-SysAdmin mailing list
[email protected]
http://hcoop.net/cgi-bin/mailman/listinfo/hcoop-sysadmin

Reply via email to