quota

Jeffrey A. St. Pierre Jasp2 at Colorado.EDU
Thu Jul 15 16:46:55 UTC 2004


On Thu, 15 Jul 2004, Ryan Golhar wrote:

> I'm not sure that will do it.  From my understanding, quotacheck reads
> all the files and creates the quota file.  When you run quotacheck, then
> partition should be unmounted so as to prevent anyone from using it.
>
> Simply turning off quota's won't be that.
>
> I'm in the same situation, however I'm unable to unmount the filesystem
> so instead I just use -f to force it to update the quota file...
>
>

it works... and is probably less dangerous than the -f

try this:

   quotaon -ap

That will list whether quotas are on or off.

   quotaoff -a

that will turn all quotas off.  You can just specify -u or -g for 
user and groups and you can specify the filesystem if you want.. 
like so:

   quotaoff -u /var/spool/mail

Then you can run your quotacheck without it complaining about 
needing the -f force.

   quotacheck -vu /var/spool/mail

of course, while it runs this check /var/spool/mail is read-only, 
unless you specify the -m too.

When done, just turn things back on.

   quotaon -u /var/spool/mail

I tell you it works... and it's got to be better than forcing. 
Try it.

Of course this is not an issue with Solaris or BSD... just run 
quotacheck whenever you want!

-Jeff

>
> Well,
>
> I haven't recieved much of an answer from anyone on 3 different
> lists.  I guess noone uses quota that much or they just force it.
>
> Here's what I figured out though, basically write a script that does
> this:
>
> quotaoff -a
> quotacheck -agum
> quotaon -a
>
> running by hand you can add (-v) for verbosity.
> the (-m) on the quotacheck is nice, as it does not make the filesystems
> read-only while checking.  People get angry when they can't write to
> their home directory.  Even with quotas off the
> quotacheck remounts the filesystem ro unless you specify -m.
>
> I assume this is the proper way to use quotacheck, despite not much in
> the way of documentation for this.  Google shows many people asking
> whether -f is bad, but noone has the answer.
>
> Hope this helps somone, or prompts someone with real info to
> write in.
>
> -Jeff
>





More information about the redhat-list mailing list