How to solve this

Pre-Requisites : You should have Shell access (as root) to the Node server.

Mostly this happens due to inode issue. The inodes allocated to the server maybe full. inode is a data structure on a traditional Unix-style file system such as UFS. An inode stores basic information about a regular file, directory, or other file system object.

Steps

Say we are logged into the Node Server. There we are going to list all the Containers existing on the Hardware Node. By default, only running Containers are shown.

[root@linuxvps1 ~]# vzlist -a
CTID    NPROC STATUS   IP_ADDR        HOSTNAME
100     18  running    100.20.11.111  WEB.EXAMPLE.COM
101     63  running    100.20.12.203  -
102     169 running    100.20.92.182  server2.example.com
103     81  running    100.20.56.169  example.net
78965   236 running    100.20.77.204  server1.example.com

Assume your container is 103. So simply enter into 103 by this command

[root@linuxvps1 ~]# vzctl enter 103

Check if the inodes are full in the container

[root@linuxvps1 ~]# df -i

Most Probably it should display something like this :

Filesystem      Inodes      IUsed        IFree     IUse%      Mounted on
/dev/vzfs       200000      200000       0         100%       /
none            65536       95           65441     1%         /dev

Bang ! The inodes on /dev/vzfs mounted on / are full. You’ve to go nowhere else to see why that error was delivered. A simple solution will work out, you’ve to increase the inode limit. You’ve two options, either through Shell or through the Node Control Panel (like Parallels Infrastructure Manager)

First we will go through Shell mode with this command :

$ vzctl set veid --diskinodes softlimit:hardlimit 

Where ‘veid’ is the VPS ID

root@linuxvps1 ~]# vzctl set 103 --diskinodes 7680100:7680200

Second we are going for Node Control Panel. In the infrastructure, you will see your VPS with a ! sign. That indicates your VPS is having some sort of issue.

So you’ve seen the Alert. Next access the VPS, go to Resources tab. I’m sorry to tell you that I grabbed this screenshot after resolving the issue, but it will help you anyway

You should see the ‘Disk Inodes’ is full under ‘Disk Usage’ category (Here it is not !). Next Click Configure to settle this one

See the Highlighted part, the inodes should be limited to some size. Increase it or remove the check mark to make it unlimited.  Make sure you save the changes. Now try again accessing cpanel/WHM/Webmail. It should load with out any issues.That makes a smile in your face, isn’t it ?

SERVICE CURRENTLY NOT AVAILABLE!

Error No. [0x01F4]

That is what you will see on logging to the Roundcube. How to fix it ?

on cPanel, just update roundcube

/usr/local/cpanel/bin/update-roundcube --force

On Lxadmin, just re-install

rm /var/cache/lxadmin/lxwebmail*
lphp.exe ../bin/misc/installRoundCube.php
/script/upcp

Fatal error: Cannot use string offset as an array in /usr/local/cpanel/base/horde/lib/Horde/Block/Layout/Manager.php on line 389

Not sure what caused the issue and it is not happening with IE, and only in FF. However a simple layout reset fixed the issue. Good Luck. Still, we need to find the real cause.

No one was able to login to Plesk’s default webmail, horde. Enabled error reporting in php.ini and got the error message as below in the “/etc/httpd/logs/error_log” and/or “/var/log/psa-horde/psa-horde.log”

PHP Warning: Unknown(): Failed to write session data (files). Please verify that the current setting of session.save_path is correct (/usr/local/psa/admin/sessions) in Unknown on line 0, referer: http://webmail.supportsages.com/test.php?mode=unregister

Here the solution was to edit the php.ini to use /tmp as the session.save_path or even “/var/lib/php/session”

Other solution, for similar issue (not this exact error though) is to set,

session.auto_start = 0 in php.ini
Making sure that localhost entry is in /etc/hosts and resolves. For that do a telnet localhost 143 and if you get “Connected to” message, its not that issue check /etc/hosts file. It must have permissions -rw-r–r– (644) on a ls -l
Whatever be the session.save_path setting in the php.ini, it should have the permission of 1777
Run IMP test to check the problem at: http://webmail.DOMAIN.COM/horde/imp/test.php and make sure that you get the green words for all the tests.
This problem may occur if the domain’s web content is hosted on another server and the MX DNS record doesn’t exist or points not to the Plesk server. IMP tries to check mail on whatever server the domain name resolves to. In order to prevent this from happening the domain name or MX record (if it exists) must be resolved locally to the server.

References : http://kb.parallels.com/article_41_260_en.html