[BlueOnyx:17742] Re: PHP errors on Disk Usage

[Michael Stauber]

Hi Robert,

> I have these errors when trying to access disk usage under server 
> management on a 5208R VM on a CentOS7/KVM host. Disk usage works fine 
> under virtual sites. I found some posts regarding this, but no solution. 
> I have done all Yum updates, but not rebooted, logged out and back in, 
> the issue remains....
> 
> A PHP Error was encountered
> 
> Message: Undefined index: site2
> 
> Filename: controllers/serverDiskUsage.php

Many thanks for reporting this. The problem is twofold:

I just published an updated base-disk for 5207R/5208R/5209R which will
prevent from this error being raised again. But the existence of this
error indicates a problem on your server that you need to look into. It
should be easy to fix, though:

1.) Check permissions on /usr/sausalito/bin/ccewrap. They should look
like this:

[root at 5209r buildarea_5200r]# ls -la /usr/sausalito/bin/ccewrap
-rwsr-x--- 1 root admserv 64955  5. Mär 16:28 /usr/sausalito/bin/ccewrap

As you can see, ccewrap has the SUID bit set to allow it to perform
transactions with root rights. The GUI needs this for many things and it
allows the GUI (after permission and ACL checks) to run a limited subset
of commands with root rights. This is safe.

If the SUID bit is missing and the rights look differently than shown
above, then please run this command:

yum reinstall sausalito-cce-client

2.) If the permissions were correct to begin with and/or the GUI now
shows vsites without quota information, then please run this command as
well:

/usr/sausalito/sbin/get_quotas.pl --sites --sort=name descending

This command is used by the GUI page controllers/serverDiskUsage.php to
check the disk quota and usage of Vsites. Normally it should produce
results similar to this:

[root at 5209r buildarea_5200r]# /usr/sausalito/sbin/get_quotas.pl --sites
--sort=name descending
site1   19304   512000
site2   19264   512000
site3   18912   512000
site4   18908   512000

So the output are three columns. Column #1 for the group name (siteX),
column #2 for the actual disk usage in blocks and column #3 for the
allowed quota (in blocks).

If your output of this command looks drastically different and doesn't
follow this format (or lists all sites!), then please run this to clear
this up:

/usr/sausalito/sbin/disk_restorequotas.pl
/usr/sbin/swatch

That should fix it.

-- 
With best regards

Michael Stauber