Plesk 11.x / Linux :: All Sites Unresponsive Until Reboot Happening Several Times A Day?
Jun 8, 2014
OSCentOS 5.6 (Final)
Panel version11.0.9 Update #61
The system is up-to-date; last checked at June 6, 2014 03:32 AM
I use pingdom to alert me when sites go down. It would happen only once every few weeks, and I'd just reboot the VPS from my provider's control panel, however in the last 24 hours it has started doing it more and more often. The only change I have made in this time is to delete a few unused web sites.
When I get alerted, I check the the domains and they won't load. I look at the plesk control panel page thought and that does work.
I can SSH in, but I can't see anything that might work in the messages log.
I removed the plesk control panel error as that only seemed to happen once and is not happening anymore.
I saw quite a few entries in the /var/log/secure log and so I added iptables rules to only allow SSH access from my own IP and to DROP all other requests. It seemed fine for a while, but after 12 hours it has started becoming unresponsive again until reboot.
I run a centos 6 plesk 12 server with around 7 domains.All runs perfect untill a 3 hours ago. The server was offline.After reboot plesk starts up without a problem but the sites keep offline. While httpd and mysqld and psa runs perfect.Also cannot find error messages in the messages, error_log or anything. How to proceed debugging?
Upon a shutdown or reboot, the system shuts down sw-collectd. Further along, it will end wdcollect and the following will occur:
init: psa-wdcollect main process (pid) killed by KILL signal init: psa-wdcollect main process ended, respawning wdcollect[pid]: Language en-US is used for sending e-mail messages. wdcollect[pid]: Failed to connect to database server during the startup. New attempts will be made if needed. wdcollect[pid]: Server started
I believe this is preventing the un-mounting of drives which in the end freezes the shutdown process on:
Please stand by while rebooting the system...
After this occurs, I have to force the VM off and then boot again.
When I reboot my centOS 6.5 + Plesk 11.5 server, Nginx doesn't start and every time I need to go to Server > IP Addresses and click Reread IP.Then Plesk says that need to repair one of these ips, I click repair and then Nginx can start. ( I have 3 ips [2 ipv4 and 1 ipv6] )
So Everything was going great! Until I needed to reboot, and then after that only Plesk panel worked not my IP or any domains setup with it.
BUT when I restart Nginx it all starts working again! (Had to /usr/local/psa/admin/bin/httpdmng --reconfigure-all to get the domains to go to the right directory though)
But now everytime I reboot the server Nginx has to be restarted
Two month ago I installed plesk 11.5 on my server (with centos 6 or 7 I don't remember) Then i added my domain.com by clicking on "Add New Domain" and using the auto dns system configuration. All worked great. During this period, Parallels Plesk was auto updating to 12.x 1 week ago, I restarted my server, and my website was down (I think because DNS problem with 12.x or ip problem or port closed...)
So I formatted my dedicated server, then I reinstall Parallels Plesk 11 (with centos 6) and the website worked great. Then I re-formatted my dedicated server, then I installed Parallels Plesk 12 (with centos 7) and the website did not work.
So I think the problem come from plesk 12
This is how I process to configure my website on plesk 12 :
1. This is a 1&1 domain and an OVH dedicated server.
So I configure DNS like that :
1rst name server : nsXXXXXX.eu 2nd name server : sdns2.ovh.net
2. On plesk, i click on "Add New Domain"
Capture : [URL] ...
3. I wait for DNS propagation but it doesn't work...
Recently upgraded to Plesk 11.5 & upgraded to the latest CentOS kernel. Whenever I reboot the server DNS services do not work, can't view any websites hosted on the server etc.
Once I either restart BIND through Plesk Panel or
Code: # service named restart
Everything works perfectly.
My var/log/boot.log shows named started [OK]
When I run
Code: # service named status
Immediately after reboot everything looks normal, despite not resolving DNS until restarted.
I ran
Code: # ceckconfig --list
And both httpd & named services are displaying all 'off' on every runlevel. But, obviously named service is starting automatically, maybe by psa?
Should httpd & named be listed as on, is this the problem?
Started after last update 12.0.18 Update #44 (on CloudLinux): I can send mail wia created mailboxes, but sites can not send mail, contact forms say "Failed to send your message." maillog file also do not indicate sending any messages from domains. Wordpress restore password claims:
"The e-mail could not be sent. Possible reason: your host may have disabled the mail() function." php.ini is default: safe_mode = Off disable_functions = is empty
The issue about not sending mail from domains was - old versions of LVE anf CageFS are incompatible with latest plesk update, updating them caused all sites in Cagefs to throw error 500 with log string: "SecureLVE jail error Invalid regexps: no match for path to base home directory". It was fixed by changing cagefs.base.home.dirs regexp to the real sites folder location. Enabling cagefs back removed one of the files it shouldn't:
"Removed file /var/cagefs/81/xxx/etc/php.d/soap.ini" So be carefull to place necesssary files back. For now everything works. Click to expand... Click for original message
I just want to remove plesk and installed again. i saw (URL...) this page. use this code but plesk not removed. my sites are gone.how can i get back my site? what i should install.When i try install plesk 12 says you already install version 11.0.9
We have created symlinks, modified some configuration files (Plesk included) and ran domains reconfiguration script (zero domain installed, but just in case the script will do anything else).
After that, we have launched Plesk Migration Wizard (from Plesk 9 to last release).
Everything has worked fine: all domains have been migrated on our new folders (/opt based) and owners has been correctly defined for each domain. The same for MySQL and all mail accounts.
We have checked Apache VirtualHost configured by Plesk Migration Wizard : our new folders are used (/opt).
So all seems to be ok.
But now, all websites (configured with Apache and Apache PHP module) returns 403 Forbidden error.
We have try changing to FastCGI with Apache same error.
We have try changing to full Nginx (PHP and static files), some websites run.
We have checked some KB (112884, 800) but all seems to be good.
Is there any steps we have omitted when we move folders ?
After successfully changing to a fresh created dhparam pem and a reissue of my certificate all was well.
A couple of other things needed to be done so I followed the article: [URL] ....
nginxDomainVirtualHost.php was already present so the only thing I did was adding the ciphers I got from another site (ciphers that also gave me XP and IE8 support etc.)
After executing the httpdmng --reconfigure-all command I instantly got a error message:
I am seeing some some some strange behaviour when password protecting directories served by nginx and PHP-FPM. If I have a site set up so that 'Process PHP by nginx' is selected under ('Websites & Domains>Web Server Settings>nginx settings') and set up password protection ('Websites & Domains>Password-Protected Directories') PHP pages are still served without asking for a password.
If I untick 'Process PHP by nginx' the behaviour returns to normal and an attempt to any access files results in the password request.Is this behaviour by design? If so, it is not made clear when you set up the password protection that it will not apply to PHP pages if you have nginx process the PHP pages.
Recently, we had trouble with Horde, which I Re-installed. However, since then any change that is made to a domain in plesk (changing mailbox settings, domain settings, hosting settings) All website show the default plesk page and I will have to re-run the reconfigure-all command in SSH to bring them all back up. Sometimes even that won't work and I will have to re-run the bootstrap repair command. We were actually on Plesk 11 when this started happening, so we upgraded to Plesk 12, but unfortunately this did not fix the issue.
We have around 500 customer websites on this server, so obviously the changes are made often and the sites going down on every change isn't great at all.
I have a long list of these errors in my error Plesk error log:
2014/09/25 17:31:53 [error] 818#0: *1566 FastCGI sent in stderr: "Primary script unknown" while reading response header from upstream, client: *.*.*.*, server: , request: "$
But I can't seem to find out if this is the cause, or how to fix this error correctly on a Plesk install.
I am trying to increase the PHP memory limit for all Plesk hosted sites.
I've updated the 'memory_limit' setting in /etc/php.ini
Then I ran
Code: /usr/local/psa/admin/sbin/httpdmng --reconfigure-all /usr/local/psa/admin/sbin/httpdmng --reconfigure-server Checking a phpinfo() page on hosted sites I see that this setting has not been effective.
Checking some /var/www/vhosts/system/domain.com/etc/php.ini also shows the old setting, and I can see that these files have not been updated.
I note that all these domain php.ini files have been modified earlier today, all at the same time, so presumably some process did regenerate them earlier. All these files hae the "DO NOT MODIFY ..." heading
I realise that I can go the the domain in Plesk GUI and save the PHP settings to force an update, but is there anyway to force the regeneration of all of these files via CLI?
When I turn on nginx for the server I am able to browse non SSL protected sites correctly. I receive a 502 Bad Gateway error on any sites that have an SSL certificate. Is there a fix for this? I have been unable to find one.
I have a new DS with Plesk 12. I needed to change the default /var/www/vhosts location to a different physical drive, with more capacity.The larger disk is mounted as /disk1, so I created a folder here called vhosts, moved everything from the existing /var/www/vhosts directory, then I mounted /disk1/vhosts as /var/www/vhosts. See my fstab output below.
I then edited /etc/psa/psa.conf to update the VHOSTS location as per [URL] ...., and restarted Apache, MySQL and reconfigured domains.I thought all was OK, and that Plesk would just see /var/www/vhosts as normal, but I proceeded to migrate domains from another server to this box, and found that whilst some sites were working OK, many were not.
I was finding that existing config directives that used <Directory> were being ignored or seemed to disappear.All sites with .htaccess files stopped workingWe would get Apache / PHP errors saying that files could not be accessed as they were not in the allowed path for the domain. Plesk seems to be logging, reporting, and handling files as if they were in /disk1/vhosts, not the usual location. I have got sites working by adding /var/www/vhosts/ domain and /disk1/vhosts/domain to the PHP Settings page as allowed paths.
I have checked in the httpd.conf for each domain, and they all list the site document root as /disk1/vhosts/domain, not /var/www/vhosts/domain.It's as if Plesk is ignoring the change in /etc/psa/psa.conf.
All I want is for Plesk to think that /var/www/vhosts is the document root for all domains, and not do anything with /disk1/vhosts. I have many sites with hard coded links to included files in /var/www/vhosts/domain, and I dont want to have to rewrite them all to use /disk1/vhosts.
I have had to stick HTTP configuration directives on nearly every site, and add these entries to the allowed paths box, which obviously I should not have to do. Can any Plesk / Linux gurus tell me what I need to do to fix my problem, get sites working in /var/www/vhosts, and have Plesk allow .htaccess files to work without me having to add "AllowOverride All" to every domains??
# /etc/fstab # Created by anaconda on Tue Oct 14 05:22:57 2014 # # Accessible filesystems, by reference, are maintained under '/dev/disk' # See man pages fstab(5), findfs(8), mount(8) and/or blkid(8) for more info #
I have a Plesk 12 server running that works great. But the problem is that whenever one of my shared hosting customers changes something out a domain or subdomain, all the other sites on the server give a 502 error.
The reason why is clear to me: apache is reloading its config while nginx is still active, but sees apache as being down and therefor throws a 502 error.
Is there any way to prevent this from happening? Because it's happening a lot and people are sending in 'site down' reports while there's actually nothing wrong with the server or their site.
For some reason I am having issues with latest Plesk with php and wordpress. HTML sites are fine, but all php and wordpress sites are giving a 502 error after fully migrating. I repaired with Webserver Configurations Troubleshooter and rebooted and its still giving me the same issue.Also, noticed my previous server had php 5.4.36, the new CentOS server has 5.4.37.
I recently purchased a Dedicated Server, and this is the first time I have ever used a Dedicated Server before, although I do have some experience with Linux. The problem I am having is that if you assign an Unused Dedicated IP to a domain using WHM the entire website that is assigned to that IP becomes inaccessible. I am able to ping the Dedicated IP, but that's about it. So the problem is probably that the Dedicated IP isn't pointing to the proper location where the website is, if you use the Server IP everything works fine.
So my question is what do I have to do to fix the problem and point the Dedicated IP to the proper location? Is it something that the Host has to do for you or is it just a simple file edit as root? I am using CentOS 5.3 64-bit as my Operating System with cPanel/WHM as my Control Panel.
