Plesk 12.x / Linux :: Wrong Calculation Of Disk Usage
Mar 12, 2015
I'm running Plesk 12.0.18 on Centos 7, recently I've got emails from backup service stating that it could not complete the backup due to insufficient space available on disk. Normally I have plenty of disk space available so I check the disk using the command
df -h
Filesystem Size Used Avail Use% Mounted on
/dev/root 20G 2.3G 17G 13% /
/dev/md2 91G 82G 4.4G 95% /var
it seems that /var has been growing up a lot,
but if I run the command
du -sh /var
I get a total size of 5.7G (not 82G as stated before)
is Plesk calculating the wrong size or it's me using the wrong commands?
I migrated from one server running Plesk 10.4.4 Update #59 to a new server running Plesk 11 which I then updated to 12.0.18 Update #32. Since moving to Plesk 12 the "Notify when disk space/traffic usage reaches" emails have been being sent to our customers set up on the server whereas previously they were sent to the server admin email address.
I'm on PPA 11.5 MU#2 (Should upgrade to MU#3 soon). My problem is that the Usage of Disk Space for all my customers subscriptions are not calculated. I've run the daily maintenance script (which it actually runs periodically), but there's no update on the display.
I recently set up an eas and autodiscover sub-domains on my main domain (eg myserver.com) running Plesk, Webmail etc, and is the default website for the ip.
So it should be..
myserver.com (not SNI default for IP) eas.myserver.com (SNI) autodiscover.myserver.com (SNI)
But I notice the actual default site (non SNI) is the sub-domain autodiscover.myserver.com.
I checked in
Plesk -> Tools & Settings -> IP Addresses
and that was set correctly, so I tried to change it and back again, then...
/opt/psa/admin/bin/httpdmng --reconfigure-all
Neither worked.
I'm running Ubuntu 12.04.5 LTS and Plesk 11.5.30 #48 (the latest MU)
I had a old virtual Server with Ubuntu+Plesk12. I build a new dedicated Server with CentOS6+Plesk12 and migrate all Domains. Now i have a strange problem:
When i upload a file with FTP-User "user1", the file has the permissions Group=psacln & Owner=user1. All is fine, i can see and edit this file in a FTP-Program. But when my website create a file (cache html files or installed plugin folders in wordpress), I can't see, can't edit and can't download this files.
When i login to plesk with the admin account, i can see and edit this files. They have also Group=psacln & Owner=user1 and i can give them 0666 rights but when i login with FTP-User "user1" with my FTP-Program, I didn't see that files?!
I'm new to Plesk - I've inherited a web site for a local social club so have just been handed login and password. I believe I have version 11.0.9 and as I can see mention of an Apache webserver I'm guessing its linux.
So - I want to redesign and use Joomla so I can share the content updating with others.
I can't install Joomla as I have PHP 5.1 and need 5.3. I have no 'server' or any sort of option I can find (and I've looked everywhere!) to find a command line or somewhere I can run things.
When I try to allow SSH access the option says Forbidden and there is no option to change it.
I have a problem where every folder I ftp onto the server gets given a 700 permission and every file gets a Zero permission. Most of the sites I am adding to this server will be WordPress sites and most of the folders I want to have a permission of 755 and for the files a permission of 644 so my question is how do I change the settings so that any new files ftp'd onto the server get the permission 755 and files get the permission of 644 ?
I set this up for someone, and temporarily changed the email for the administrative account to my address.
I have since changed this back to the original email, but Plesk keeps sending the administrative emails to my account.
I've pretty much grepped through the entire server for my email in order to stop it sending me emails, but I can't for the life of me find out why it keeps sending these emails to me.
How do I make Plesk send these emails to the right email-account?
I have noticed a weird thing lately with two VPS accounts. Virtuozzo's system usage meter on the Resources page creeps up to 100% over the course of 24 hours or so and stays there, but stays green and I don't get any QoS alerts. The servers both seem to be running fine, and I haven't had any crashes. One VPS is running a few Wordpress blogs.
The other is running some Joomla sites.
Both are running Virtuozzo on CentOS 4.7 i686 with WHM 11.24.2 and cPanel 11.24.4-R32486.
In the past, when the system usage meter in Virtuozzo got anywhere near 100%, it turned yellow or red, I had sites go down, and all kinds of QoS alerts for privmpages were thrown. It turned out back then that I had a bad PHP script eating tons of memory with too many MySQL queries. That doesn't appear to be the case here, since I'm not having any apparent problems with either server. It seems like a "non-problem" but it has me concerned because I'm used to using that system usage meter in Virtuozzo to keep an eye on the health of the VPS..
When I look at memory used in WHM (WHM->Server Status->Server Status) it shows only 31% used, which is what I used to see in Virtuozzo.
When I run free, I get:
root@server [~]# free total used free shared buffers cached Mem: 524288 133804 390484 0 0 0 -/+ buffers/cache: 133804 390484 Swap: 0 0 0 When I run Free, I get: root@server [~]# Free Memory limit...: 8388604 MB (burstable: 8388604 MB) Current usage..: 120 MB (0%) (requested: 264 MB) Free memory....: 8388484 MB (100%)
Initially, I suspected it was another problem with MySQL, and restarting MySQL does drop the system usage meter down to ~31% in Virtuozzo for a while--without changing the actual memory used on the System Processes page--but the memory usage for MySQL with Virtuozzo pegged is only 6.23% in WHM.
Here's the really weird thing: If I look at Virtuozzo->System Processes it shows only 0.4% (I think they mean 40%) of total memory used, half of which is mysqld and the rest is spamd.
With no sign of problems, it looks like Virtuozzo is incorrectly reporting memory usage in the system usage meter, but correctly in the system processes list.
The web host's response was:
We use SLM for managing VPSes. This enforces your per VPS limits and doesn't let your VPS use more then it should be. It sounds like you guys may have some memory hogs in your applications.
Have you guys considered cutting down some of the unneeded plugins and stuff?
That doesn't make any sense to me given the disparity between the memory usage meter and the amount of memory usage reported in System Processes and with free/Free as well as the lack of QoS alerts, but maybe SLM prevents any alerts? I don't want to let a memory hog run free on my system if one exists.
the hostname is www.domain.com also in the serversettings of plesk. When we go to http://www.domain.com:8443 we become redirected to https://www:8443. The only way to connect to plesk is to use https the redirect from http doesnt work.
i'm running the latest version of Plesk 11 on a Ubuntu 12.04 system.
We have a customer with a domain and this customer added other domains to his account.
now, 2 domains are not working, he created the as usual, Plesk created the directories under /vhosts/domain.com/domain1.com and the vhosts.conf are also correct.
when i'm opening the domain in the browser, i get the following error message:
The requested URL /var/www/vhosts/domain.com/index.php was not found on this server.
I have a problem with webmail (horde). Horde Webmail show the wrong time, 2 hours to late. It look like that horde show the UTC time, but not the CEST "Europe/Berlin" of my server. I can change that over "preferences" -> "global preferences" -> "locale and time" and change "Your current time zone" from "Default" to "Europe/Berlin". This will now show the correct time (timezone), but this is not the solution because this will do all webmail users.
How can i change the default timezone in Horde?
Output from shell
#my timezone and this will show in all things but not in webmail horde [root@www /]# date Fr 11. Jul 17:32:14 CEST 2014
#this time look like the default time in horde [root@www /]# date -u Fr 11. Jul 15:33:31 UTC 2014
Seems this started when upgrading to a version of 12. It was working a few days ago and only seems to affect mailing lists. I found a google Cached thread here where Igor was assisting some folks as late as Aug 4 and referenced this was "reported to development (PPP-10678 for your reference)" it seems the forums on Parallels changed or something because several Google links are not working and resulting in having to used cached results for the two links below.
I am able to disable SPF and the e-mails go through just fine however this was working with SPF enabled before a recent upgrade.
Page 1 [URL] ....
Page 2 [URL] ....
This is the error message displayed in /var/log/maillog Sep 24 01:38:35 controlpanel postfix/smtpd[3725]: connect from localhost[127.0.0.1] Sep 24 01:38:35 controlpanel postfix/smtpd[3725]: D565017C013E: client=localhost[127.0.0.1] Sep 24 01:38:35 controlpanel greylisting filter[3899]: Starting greylisting filter...
apparently Envelope-From, Sender and Errors-To headers are set to a wrong address for messages from mailman to the list-owner. This is particularly true for mails like 'message requires approval'.
It seems like the plesk<->mailman magic results in all list-domains being written to the end of /var/lib/mailman/Mailman/mm_cfg.py in the following format: add_virtualhost('list.DOMAIN').This in return results in mail headers Envelope-From, Sender and Errors-To of mailman system mails like described above being set to mailman-bounces@DOMAIN, where DOMAIN is the one from the last add_virtualhost entry.
In my opinion, this obviously is a bug. Maintainance mails from mailinglists have wrong headers set, often with a domain that belongs to a completely different customer than the mailinglist.
I found the following discussion on mailman-users mailinglist from 2008 which discusses the same issue - also a plesk setup: URL....
We're trying to migrate a Webspace from a 11.5 to a 12.0.18 Plesk host using the Migration Manager. The Webspace contains 4 domains. The main domain of the webspace gets migrated fine, however, all of the "child" domains fail with the following error:
Stderr is Wrong syntax for command's '-webspace-name' parameter. 'webspace-domain.ch' does not match the pattern '/^[[:alnum:]]([[:alnum:]-]{0,61}[[:alnum:]]){0,1}(.[[:alnum:]]([[:alnum:]-]{0,61}[[:alnum:]]){0,1})+$/iu'.
I am getting problem , my package is 500MB and i have checked cpanel > file manager , there is showing disk usage 350 MB but there is no file under public_html/ and out of public_html/ directory , what can i do and why this is showing 350 MB disk usage but there is no files.
whm only shows this in server information area > disk usage
Code: Filesystem Size Used Avail Use% Mounted on /usr/tmpDSK 243M 4.4M 226M 2% /tmp /tmp 243M 4.4M 226M 2% /var/tmp It is not showing actual disk space for sda1/sdb1
I restarted crond and that did not fix it, nor forced cpanel update.
I cannot trace where the usage is coming from. There is only one account on this box using 26GB. Its a centos/cpanel box. I checked /var/log and had already deleted audit.d directory. /usr/local/apache/domlogs and logs show almost nothing.
the above commands done nothing - really waste of time!
The cPanel asked my to contact my VPS provider and ask them to reinitialize quotas for your VE and possibly check further into the node to correct the issue. As for VPS provider they did some tests & told me that they "fixed" & can't find any issues on the node, but the problem is still exists.
asked by VPS provider to do: /scripts/fixquotas restart VPS /scripts/upcp --force
how to fix cPanel bug?
The attached image are proof of Disk usage not being updated in WHM & cPanel, as this account contains 17.6 MB (18,472,960 bytes).