How to increase Import file size limit in PHPMyAdmin

One thing a lot of people do is to change the upload_max_filesize in their php.ini and go on their way. While this does change the maximum upload file size, this doesn’t really change it when using PHPMyAdmin

You should change all three of these:
memory_limit
post_max_size
upload_max_filesize

Then once you do, restart your web server

If you are running Apache
sudo /etc/init.d/apache2 restart
or
service httpd restart

if you are running IIS
iisreset /restart

If you are in a shared hosting and do not have root access or access to your php.ini configuration file, then the other way is to create a .htaccess file in your application root and add the following lines (replace the 25M with the value you want to use).

php_value upload_max_filesize 25M
php_value post_max_size 25M

You do not need to restart your web server when using the .htaccess file, but please note that this will really only work on Apache (windows or Linux)

If you need .htaccess to work on IIS you may need to look at a third party product like http://www.helicontech.com/ape/

Scheduled Tasks / Cronjobs Not Running Since Upgrade to PHP 5.3 / MySQL 5.5

Since upgrading PHP to 5.3 and the latest MySQL 5.5, our Scheduled Tasks / Cron jobs have not been running.

Trying to run the backup manager manually I see the following:

# /usr/local/psa/admin/bin/backupmng
backupmng: Unable to fetch backup tasks from database: BIGINT UNSIGNED value is out of range in ‘((-(weekday(`psa`.`BackupsScheduled`.`last`)) + `psa`.`BackupsScheduled`.`backup_day`) – 1)’ backupmng: Unable to fetch backup tasks from database: BIGINT UNSIGNED value is out of range in ‘((-(weekday(`psa`.`BackupsScheduled`.`last`)) + `psa`.`BackupsScheduled`.`backup_day`) – 1)’

After some searching and trial and error it is actually an incompatibility between backupmng and MySQL 5.5.   To prevent unsigned int overflow with MySQL, alter the table so the BIGINT is now just a normal INT.

ALTER TABLE BackupsScheduled MODIFY COLUMN backup_day INTEGER;

You can run this command right from the normal CLI interface if you prefer like this

mysql -uadmin -p`cat /etc/psa/.psa.shadow ` psa -e “ALTER TABLE BackupsScheduled MODIFY COLUMN backup_day INTEGER;”

That’s it – your backups should now be able to be run both manually and from the normal cron tasks.

Unable to init SSLCertificate object: SSLCertificate: unable to select: no such row in the table

Every so often I get an error like this when attempting to remove a domain name or a client who at one either had an SSL Certificate in Plesk, or tried to set one up and it never really got there.

Here is how to fix that.

I just wanted to reply that this is in fact the answer – I ran into the same problem today and used these steps to fix.

go into the server via SSH and enter MySQL

mysql -uadmin -p`cat /etc/psa/.psa.shadow ` psa

Then use the query to find the domain name that is bad

Code:
SELECT d.id,
       d.name,
       R.rep_id,
       R.component_id,
       c.id
FROM   domains d,
       Repository R
       LEFT JOIN certificates c
              ON c.id = R.component_id
WHERE  d.cert_rep_id = R.rep_id
       AND c.id IS NULL;

The response will look something like this, although exact result of yours will be a little different

+—–+————-+——–+————–+——+
| id  | name        | rep_id | component_id | id   |
+—–+————-+——–+————–+——+
| 505 | domain.com |     27 |           16 | NULL |
| 505 | domain.com |     27 |           25 | NULL |
+—–+————-+——–+————–+——+
2 rows in set (0.00 sec)

Then update the database and you can then remove the domain

 update domains set cert_rep_id = NULL where id=505;
Query OK, 1 row affected (0.01 sec) Rows matched: 1  Changed: 1  Warnings: 0

Clicking on Custom View Settings in Tools throws exception

Repro Steps
Under Server Management left nav item, click on the Tools and Settings
Under Panel, click on the first item -> Custom View Settings

Immediately an exception is thrown

PHP Catchable Fatal Error: Argument 1 passed to CommonPanel_Controller_Helper_Subscription::getPer missions() must implement interface Plan_Interface, null given, called in /usr/local/psa/admin/panel/application/library/AdminPanel/Form/Final/ServicePlan/AdminPlanHostingPermissions.php on line 45 and defined
Line: 167
File: /usr/local/psa/admin/plib/CommonPanel/Controller/Helper/Subscription.php
Stack Trace:
[ The stack trace is too large to post here, it is at http://pastebin.com/rXaSypcp ]

My Interface Views are set to Service Provider view, but even if that is the cause it should not throw an exception, it should either hide this area from being accessed, or provide a friendly error message. Well here is how to fix that:

Basically you go into SSH and run these 2 commands, and it will be fixed afterwards:

# /usr/local/psa/bin/sw-engine-pleskrun /usr/local/psa/admin/plib/scripts/create_default_simple_template.php
# /usr/local/psa/bin/sw-engine-pleskrun /usr/local/psa/admin/plib/scripts/upgrade_admin_simple_plan_permissions_1013.php
#

Nginx breaks SSL certificates chain

SSL certificates for certain domains are not trusted anymore.
In fact any SSL certificate installed into Parallels Plesk Panel (PP) 11 with more than a single chain certificate on it is unable to load properly.
The only way to fix it is to switch back to apache using “/usr/local/psa/admin/bin/nginxmng –disable” command.

Apache has a ‘SSLVerifyDepth’ parameter with a default value of ten (10), which means that it will look for ten (10) CA (Chain) certificates. Nginx has a similar parameter ‘ssl_verify_depth’, but with a default value of one (1). Since Nginx is the front end web server and it is set to one (1), it is not grabbing the additional CA certificates, thus causing a conflict with some browsers.

Resolution

The issue has been fixed since version 11 MU#10, however it may still affect certificates which had been generated before the needed update was installed. In order to fix such certificates run the following commands:

# wget http://kb.parallels.com/Attachments/22393/Attachments/reload_ssl_certificate.zip
# unzip reload_ssl_certificate.zip
# php reload_ssl_certificate.php

How to Backup and Rebuild the DNS Configuration

DNS zones can get pretty screwy some times.  If they are wrong, or if named wont start for some reason, this is a good place to start.

Backup and rebuld the DNS configuration by running the below commands:

cd /var/named/run-root/etc
mv named.conf named.conf.bak
cp named.conf.default named.conf
for x in `ls`; do /usr/local/psa/admin/sbin/dnsmng --update $x; done;/etc/init.d/named restart

What version of plesk do I have? What is my microupdate version?

These days its hard to know if you are “up to date” or not.  With the patch levels that parallels uses, you could be out of date even if you use the latest version.  Here is how to tell:

cat /root/.autoinstaller/microupdates.xml

<?xml version="1.0" encoding="UTF-8" standalone="yes" ?>
<patches>
<product id="plesk" version="11.0.9" installed-at="20120718T115141">             <patch version="16" timestamp="" installed-at="20121001T040844" />
</product>
</patches>

So now you know you are at plesk version 11.0.9 and MU 16

tomcatmng is failed

This is one of the lamest messages ever.  You go to delete a domain or do just about anything and all it tells you is ‘tomcatmng is failed’.

Well, here we go.  You have the error

tomcatmng is failed –remove-service –vhost-name=domain.com

So now go to the servers error log and see what happened

tail /var/log/sw-cp-server/error_log
tomcatmng: Unable to remove /usr/share/tomcat5/psa-users/domain.com.xml: No such file or directory
System error 2: No such file or directory

Create the file and try again
touch /usr/share/tomcat5/psa-users/domain.com.xml
chown root:tomcat /usr/share/tomcat5/psa-users/domain.com.xml

Now try again and it should work.