• If you are still using CentOS 7.9, it's time to convert to Alma 8 with the free centos2alma tool by Plesk or Plesk Migrator. Please let us know your experiences or concerns in this thread:
    CentOS2Alma discussion

Fed up with Plesk

KamalG

Basic Pleskian
I have a cPanel VPS server that was installed more than 6 months ago and I hardly have any problems with it.. it's stable running and is only 1 GB, on the other hand I have a VPS with Plesk Installed on it with 2 GB memory.

I haven't done any configuration changes, no updates nothing ...
I only fresh installed it about one week ago because of too many unknown errors coming out from nowhere ..

Now I checked again I am getting 500 Internal server error and when I try to access the admin panel it gives me the following error

ERROR: Class 'Zend_Log_Exception' not found (Stream.php:81)

Search for related Knowledge Base articles

I rebooted the server and now it's loading fine ..
 
Now again

Internal error.
ERROR: Zend_Db_Statement_Exception: SQLSTATE[HY000]: General error: 1 Can't create/write to file '/tmp/#sql_c10_0.MYI' (Errcode: 23) (Pdo.php:234)

Why does these error show ?
 
Make sure that the /tmp directory is granted full permissions.

# ls -ld /tmp
drwxrwxrwt 3 root root 12288 Mar 16 22:49 /tmp

Full permissions can be granted to the /tmp directory with the following command:

# chmod 1777 /tmp
 
I already have the permission as you have mentioned.

drwxrwxrwt 3 root root 20480 Sep 16 08:08 /tmp

I don't have much traffic and the memory consumption never goes high.
Code:
# free -m
             total       used       free     shared    buffers     cached
Mem:          2150       1253        896          0          0          0
-/+ buffers/cache:       1253        896
Swap:            0          0          0

But don't know why I get 500 Internal server error and the Plesk License error sometimes when I login to the admin panel, when I restart and get back it gets fixed automatically.

I remember this morning I received email from the server which had the following

Attention,

Parallels Plesk was successfully updated.
To view the list of changes, visit http://download1.parallels.com/Ples...tm?platform=unix&locale=en-US&version=12.0.18.

This message was sent automatically by Parallels Installer.

I am not a Server guy so I don't really know where to look for.
 
Again another problem, This happened after the Plesk Updated Automatically. Should I stop the automatic update for Plesk ? My VPS is un-managed VPS and the Plesk License came pre installed with the VPS.

I can't buy a support request because not in the possession to do that. What I have experienced is Plesk is forcing users to pay for the support.

Plesk.png

Plesk version 12.0.18 Update #17, last updated at Sept 16, 2014 03:40 AM
The system is up-to-date; last checked at Sept 16, 2014 03:40 AM

Kindly check the above screenshot, the error is after I logged in this morning to check yesterday it was working fine after the reboot
 
Thank you for your reply.

cat /proc/user_beancounters > before_failure.txt
File is Empty

After restarting Apache checked this below file
cat /proc/user_beancounters > after_failure.txt
File is Empty

I don't think I will be able to use this command vzctl set <CT_ID> --save --<RESOURCE_NAME> <BARRIER>:<LIMIT>

I clicked the regenerate all configuration files and its fixed for now.
 
It might be worth checking that you aren't running out of disk space or inodes.

I'm not sure that a reboot would resolve a problem with lack of disk space unless it cleaned some files out on boot, so it is unlikely that I'm correct, but it is worth checking just the same.
 
I don't think it was because of disk space
Code:
df -h
Filesystem      Size  Used Avail Use% Mounted on
/dev/vzfs        25G  9.1G   16G  37% /
none            1.1G   24K  1.1G   1% /dev

Every time I had such errors and 500 Internal Server errors I used to restart and it used to get fixed.. I didn't generate any configuration files because they never worked..
 
I don't think it was because of disk space
Code:
df -h
Filesystem      Size  Used Avail Use% Mounted on
/dev/vzfs        25G  9.1G   16G  37% /
none            1.1G   24K  1.1G   1% /dev

Every time I had such errors and 500 Internal Server errors I used to restart and it used to get fixed.. I didn't generate any configuration files because they never worked..

Have you checked disk inodes too?
 
Have you checked disk inodes too?

Yes they only few percent used.
Code:
df -i
Filesystem      Inodes  IUsed   IFree IUse% Mounted on
/dev/vzfs      2560000 130749 2429251    6% /
none            275251     96  275155    1% /dev

Sincerely
Kamal Gurung
 
Last edited:
How to change "ulimit" and mysql "open-files-limit"

I think that the mentioned failure ( http://forum.parallels.com/showthread.php?317753-Fed-up-with-Plesk&p=753989&viewfull=1#post753989 ) resists in fact, that mysql can't have more open files than the actual "ulimit" definitions, with are defined with 1024 as a standard. Normally, this standard configuration is quite o.k., but when you have mail-server, webserver, nginx, php5-fpm and other things running on your webserver, it should be raised to be sure, that you don't reach the maximum. Some users already have higher definitions ( root and www-data ), but there is still room to adjust these settings for the user nginx, psaadm, psaserv, psacln, sw-cp-server and mysql. Be aware that each system has a global system-wide limit, which you may discover with "cat /proc/sys/fs/file-max" ... please don't change the ulimits to a higher definition, because it won't work. :p

First, you might want to check your ulimit for the current user logged in:
ulimit -n

As you now can see ( if you didn't change it ^^ ), it will display 1024 .... with the command:
ulimit -a
... you could check all limits. - Please see the man-pages for more options in ulimit.


You could do the following steps to define some higher standards:
vi /etc/security/limits.conf

you will now see something like this:
Code:
# End of file

root		hard	nofile		64000
root		hard	nofile		64000
www-data	hard	nofile		64000
www-data	hard	nofile		64000
Short explanation: You can see in the description, that you have several ways to limit some rules. There are always "4" rows in that table definition. The first row defines the user/group/wildcard ( * means all ) and you can define the range from soft to hard, The "hard limit" for file identifiers for each prozess is defined for root and www-data and because there is no "soft limit" defined, the standard of 1024 is used ... o.k. back to the possible changes. ^^

Under the previous defintions, you might now add for a pretty good standard for all users:
Code:
*		soft nofile		4096
*		hard nofile		40960
*		soft nproc		4096
*		hard nproc		40960
If you would to define the above mentioned users in detail, you would change the previous definitions as well
Code:
root		soft nofile		6400
root		hard nofile		64000
root		soft nproc		6400
root		hard nproc		64000
www-data	soft nofile		6400
www-data	hard nofile		64000
www-data	soft nproc		6400
www-data	hard nproc		64000
nginx		soft nofile		6400
nginx		hard nofile		64000
nginx		soft nproc		6400
nginx		hard nproc		64000
psaadm		soft nofile		6400
psaadm		hard nofile		64000
psaadm		soft nproc		6400
psaadm		hard nproc		64000
psaserv		soft nofile		6400
psaserv		hard nofile		64000
psaserv		soft nproc		6400
psaserv		hard nproc		64000
psacln		soft nofile		6400
psacln		hard nofile		64000
psacln		soft nproc		6400
psacln		hard nproc		64000
sw-cp-server	soft nofile		6400
sw-cp-server	hard nofile		64000
sw-cp-server	soft nproc		6400
sw-cp-server	hard nproc		64000
mysql		soft nofile		6400
mysql		hard nofile		64000
mysql		soft nproc		6400
mysql		hard nproc		64000


Now, you changed some ulimits... but actually, you sure want these limits as well in your database - server, so you have to change some config as well:
vi /etc/mysql/my.cnf

Under the definition of "[mysqld]", please add the following.

[mysqld]
open_files_limit = 6400

Now restart your mysql with:
service mysql restart
or
service mysqld restart



.... and lean back with a smile on your face and a cup of coffee. ^^ - To see the changes, you might like to logout of your shell and login in again and type "ulimit -n" to see your own user changes and if you would like to see the changes for the mysql - server, just connect to your mysql - server, using the plesk short-command:
plesk db

and use the command while your are connected to your mysql - server:
show global variables like 'open%';

... type:
exit
... to get back to your shell. :)
 
Last edited by a moderator:
Back
Top