Todays updates crashed Webmail

[ManfredZ]

Hi,
todays updates (initiated bec. of the bash-bug):

bash-4.1.2-15.el6_5.1.x86_64
php-common-5.4.33-38.el6.art.x86_64
php-cli-5.4.33-38.el6.art.x86_64
mysql-libs-5.5.40-26.el6.art.x86_64
mysql-5.5.40-26.el6.art.x86_64
php-pdo-5.4.33-38.el6.art.x86_64
1:jakarta-commons-httpclient-3.1-0.9.el6_5.x86_64
kernel-firmware-2.6.32-431.29.2.el6.noarch
kernel-2.6.32-431.29.2.el6.x86_64
axis-1.2.1-7.5.el6_5.noarch
php-mysql-5.4.33-38.el6.art.x86_64
mysql-server-5.5.40-26.el6.art.x86_64
php-devel-5.4.33-38.el6.art.x86_64
php-5.4.33-38.el6.art.x86_64
php-mbstring-5.4.33-38.el6.art.x86_64
php-gd-5.4.33-38.el6.art.x86_64
php-imap-5.4.33-38.el6.art.x86_64
php-xml-5.4.33-38.el6.art.x86_64
VirtualBox-4.3-4.3.16_95972_el6-1.x86_64
procmail-3.22-25.1.el6_5.1.x86_64
kernel-devel-2.6.32-431.29.2.el6.x86_64
tzdata-2014g-1.el6.noarch
kernel-headers-2.6.32-431.29.2.el6.x86_64
tzdata-java-2014g-1.el6.noarch
kpartx-0.4.9-72.el6_5.4.x86_64


After that, Horde Webmail stopped working when entering the Webmail lite (Calendar and rest is working)

Internal Server Error
The server encountered an internal error or misconfiguration and was unable to complete your request.

Error_log:
Premature end of script headers: index.php, referer: http://webmail.domain.com/services/portal/index.php


Any hints ?

 

[JnascECSI]

Same thing going on here also, 11.5 and all clients cannot access webmail because of the same 500 Internal Server Error.

 

[JnascECSI]

Same thing going on here also, 11.5 and all clients cannot access webmail because of the same 500 Internal Server Error.

after some more digging i found this in the http error logs, i changed the real domain to xxxxxx and i do see the same message for about 15 different domains all today.

[Fri Sep 26 09:30:20 2014] [warn] [client xxx.162.xxx.49] mod_fcgid: read data timeout in 45 seconds, referer: https://webmail.xxxxxxx.com/login.php
[Fri Sep 26 09:30:20 2014] [error] [client xxx.162.xxx.49] Premature end of script headers: index.php, referer: https://webmail.xxxxxxx.com/login.php
[Fri Sep 26 09:30:32 2014] [warn] mod_fcgid: process 14625 graceful kill fail, sending SIGKILL

 

[UFHH01]

You might want to change the config "fcgid.conf" which is located at "/etc/apache2/mods-available/" ( Ubuntu ) and add this line at the bottom:

...
FcgidMaxRequestLen 1073741824
</IfModule>

The default maximum was 1 GB, but during the times, it has changed to 128 kb. If you have no length defined, FastCGI will only process requests to a maximum of 128 kb - with this option you re-define this to 1 GB again.

For Centos and similar distributions, you find this config at "/etc/httpd/conf.d/fcgid.conf" .

 

[ManfredZ]

Hi,

this solved the 500 Internal Error page.

Now the main frame of messages and the sideframe are status "loading", but no content is showing up, also no messages in error_log.

The "New Message", and "Filter" Submenu are now working.

 

[JnascECSI]

Ok,,, That work some what like ManfredZ said but for the most part it was a dud. I did notice one thing, PSA-Horde is showing as 6.1.6 installed but when you get into a horde account it shows 5.1.5 up top and then wants you to create all of the content windows yourself. To save me the BS of phone calls i have switched everyone over to roundcube on our box which after testing about 10 accounts on separate domains seems to work like a charm. Just do not get how after running OS updates horde got mangled because that is exactly what it looks like.

I even un-installed horde all together, rebooted, re-installed, and rebooted and seems to still have the same dam issue. Only thing i can think of is that it seems to not like the latest PHP/MYSQL update because i see nothing else that has been updated in the last 4 days exept the same list that ManfredZ posted at top which fits our updates to the T and now Horde is hosed.

 

[UFHH01]

Sometimes, especially after updates/upgrades, some patches or basic packages were not completly installed.

You could always try the following commands, to get rid of most of the issues by trying the following:

/usr/local/psa/admin/bin/autoinstaller --select-product-id plesk --select-release-current --reinstall-patch --install-component base
( You can always add components to this command line, using the string "--install-component package-name" . A full list of all packages can be displayed with the command "/usr/local/psa/admin/bin/autoinstaller --select-release-current --show-components" . )

/usr/local/psa/bootstrapper/pp12.0.18-bootstrapper/bootstrapper.sh repair
( Be aware, that this path is to the ( actual ) latest bootstrapper, depending on your system and updates/upgrades/patches, this path might be slightly different. If the command doesn't work for you, try "ls /usr/local/psa/bootstrapper/" to list your own most recent version and correct the path to that command. )
​

I would suggest to investigate your logs and configuration files ( Parallels Plesk Panel for Linux services logs and configuration files ( KB: 111283 ) ) if you still experience issues/problems, because guessing might not solve issues and provided suggestions might not work.

 

[JnascECSI]

Well gave it a shot and checked the current installed components and they seem to be the right ones as you can see from below in captions. Only issue is if i try and re-install psa-horde it says it is not available with the dist, strange.

I did have another idea since my server is running the latest stuff 11.5.30 #47 and php 5.4, mysql 5.5, so on and so on. Is it possible to install/update psa-horde and all the components using the ones from 12.0.18. My thought is updating to the following packages.

psa-horde-5.2.1-cos6.build1200140821.13.noarch.rpm
psa-imp-6.2.1-cos6.build1200140818.11.noarch.rpm
psa-ingo-3.2.0-cos6.build1200140818.11.noarch.rpm
psa-kronolith-4.2.1-cos6.build1200140818.12.noarch.rpm
psa-mnemo-4.2.1-cos6.build1200140818.12.noarch.rpm
psa-passwd-5.0.1-cos6.build1200140818.11.noarch.rpm
psa-turba-4.2.1-cos6.build1200140818.11.noarch.rpm

[xxxxxxx@ecs-3 ~]# /usr/local/psa/admin/bin/autoinstaller --select-release-current --show-components
Doing restart of Parallels Installer...
File downloading products.inf3: 100% was finished.
File downloading plesk.inf3: 11%..25%..32%..44%..52%..62%..74%..80%..92%..100% was finished.
File downloading ppsmbe.inf3: 100% was finished.
File downloading sitebuilder.inf3: 100% was finished.
File downloading sso.inf3: 100% was finished.
File downloading setemplates.inf3: 100% was finished.
File downloading pp-sitebuilder.inf3: 19%..38%..58%..66%..85%..100% was finished.
File downloading billing.inf3: 15%..32%..48%..54%..70%..87%..92%..100% was finished.
File downloading mysql.inf3: 100% was finished.
File downloading apache.inf3: 100% was finished.
File downloading nginx.inf3: 100% was finished.
Checking for installed packages...
File downloading PSA_11.5.30/plesk-11.5.30-cos6-x86_64.inf3: 25%..51%..77%..87%..100% was finished.
File downloading PSA_11.5.30/plesk-patches-11.5.30-cos6-x86_64.inf3: 12%..26%..38%..44%..57%..70%..83%..91%..100% was finished.
File downloading SITEBUILDER_11.5.10/sitebuilder-11.5.10-rhall-all.inf3: 100% was finished.
File downloading BILLING_11.5.30/billing-11.5.30-rhall-all.inf3: 100% was finished.
File downloading NGINX_1.5.0/nginx-1.5.0-cos6-x86_64.inf3: 100% was finished.
Loaded plugins: fastestmirror
Loading mirror speeds from cached hostfile
* asl-4.0: www6.atomicorp.com
* atomic: www7.atomicorp.com
* base: mirror.thelinuxfix.com
* extras: mirror.pac-12.org
* tortix: www6.atomicorp.com
* tortix-kernel: www6.atomicorp.com
* updates: mirrors.adams.net

Detecting installed product components.
  common            [up2date] - Base packages of Parallels Plesk Panel
  panel             [up2date] - Parallels Plesk Panel
  bind              [up2date] - BIND DNS server support
  psa-autoinstaller [up2date] - Plesk Updater
  postgresql        [install] - PostgreSQL server support
  pmm               [up2date] - Plesk migration manager
  health-monitor    [up2date] - Health monitor
  pba               [install] - Payment gateway and registrar modules adapted to Business Manager
  sitebuilder       [up2date] - Parallels Presence Builder
  agent             [install] - Service node agent
  billing           [install] - Customer & Business Manager (recommended for shared hosting)
  mysqlgroup        [up2date] - MySQL server support
  horde             [up2date] - Horde webmail support
  roundcube         [up2date] - Roundcube webmail support
  kav               [install] - Kaspersky antivirus
  drweb             [install] - Parallels Premium antivirus
  spamassassin      [install] - SpamAssassin support
  mailman           [up2date] - Mailman mailing list manager support
  commtouch         [install] - Parallels Premium Outgoing Antispam extension
  postfix           [install] - Postfix mailserver
  qmail             [up2date] - Qmail mailserver
  msmtp             [install] - MSMTP relay only mailserver (SMTP client)
  web-hosting       [up2date] - Plesk Web hosting
  mod_fcgid         [up2date] - Apache mod_fcgid module
  mod_perl          [up2date] - Apache mod_perl module
  proftpd           [up2date] - ProFTPD ftpserver
  mod-bw            [up2date] - Bandwidth Limiter For Apache support (mod_bw)
  asp               [install] - Apache ASP support
  java              [install] - Tomcat Java Servlets support
  mod_python        [up2date] - Apache mod_python module
  cf-support        [install] - ColdFusion support for Plesk
  miva              [install] - MIVA Empresa support
  webalizer         [up2date] - Webalizer webstat
  awstats           [up2date] - AWStats webstat
  mod-cloudflare    [install] - Cloudflare Apache Module
  nginx             [install] - Nginx reverse proxy support
  phpfpm            [install] - php-fpm support for nginx (requires one of 'atomic' or 'remi' repositories enabled)
  phpgroup          [up2date] - PHP5 support
  webservers        [up2date] - Apache web server support
  psa-firewall      [up2date] - Plesk Firewall extension
  psa-vpn           [install] - Plesk VPN extension
  psa-fileserver    [install] - Plesk Fileserver extension and SMB file server package
  sshterm           [install] - SSHTerm - SSH Terminal java applet
  watchdog          [up2date] - Watchdog (System monitoring extension)
[xxxxxxx@ecs-3 ~]# yum reinstall psa-horde
Loaded plugins: fastestmirror, security
Setting up Reinstall Process
Loading mirror speeds from cached hostfile
* asl-4.0: www6.atomicorp.com
* atomic: www7.atomicorp.com
* base: mirror.thelinuxfix.com
* extras: mirror.pac-12.org
* tortix: www6.atomicorp.com
* tortix-kernel: www6.atomicorp.com
* updates: mirrors.adams.net
Installed package psa-horde-5.1.5-cos6.build115131216.11.noarch (from PSA_11_5_30-dist) not available.
Error: Nothing to do
[xxxxxxx@ecs-3 ~]#
[xxxxxxx@ecs-3 ~]# rpm -qa | egrep 'psa-horde|psa-imp|psa-ingo|psa-kronolith|psa-mimp|psa-mnemo|psa-passwd|psa-turba'
psa-mnemo-4.1.2-cos6.build115131216.11.noarch
psa-kronolith-4.1.4-cos6.build115131216.11.noarch
psa-turba-4.1.3-cos6.build115131216.11.noarch
psa-ingo-3.1.3-cos6.build115131216.11.noarch
psa-passwd-5.0.1-cos6.build115130916.13.noarch
psa-horde-5.1.5-cos6.build115131216.11.noarch
psa-imp-6.1.6-cos6.build115131216.11.noarch
[xxxxxxx@ecs-3 ~]#

 

[Namesonic]

Is there a fix for this issue? Exact same results as posted above. Webmail is unavailable.

 

[ManfredZ]

To save me the BS of phone calls i have switched everyone over to roundcube on our box

This was my first intention also, but roundcube is broken here also:

[error] File does not exist: /usr/share/psa-roundcube/services

And yes thats right, there is no services in that directory.

/usr/local/psa/admin/bin/autoinstaller --select-product-id plesk --select-release-current --reinstall-patch --install-component base

Tried this also with no effect

Please can we get back functional webmail.

 

[UFHH01]

O.k..... another try...

/usr/local/psa/admin/bin/autoinstaller --select-product-id plesk --select-release-current --reinstall-patch --install-component base --install-component horde --install-component roundcube --install-component phpgroup --install-component phpfpm --install-component mod_fcgid
​

You might ask yourself, why you should install roundcube as well, even that you just want to use horde... well the answer is, that switching from the option horde to roundcube sometimes helps to correct redirects and rewrites.

After successfull installed the above command, please rebuild the webserver configs with the command:

/usr/local/psa/admin/sbin/httpdmng --reconfigure-all
or
/usr/local/psa/admin/sbin/httpdmng --reconfigure-domain YOURDOMAIN.COM
​

Now please go to the Plesk Panel and try to switch from Horde to Roundcube in your mail - options. You then should be able to login to roundcube at: http://webmail.YOURDOMAIN.COM . If this is not possible, please investigate the error.logs either from apache2 and/or nginx and/or php5-fpm and roundcube as well ( don't forget, that there are error.logs for EACH domain as well! ) , to see, what is going wrong. The error.logs can be found at different places, please have a look at Parallels Plesk Panel for Linux services logs and configuration files ( KB: 111283 ), to find the right places for your operating system. If you wish some help with the investigations, please provide some entries of the logs, because it is much easier to suggest work-arounds or to help with misconfigurations, when people willing to help know, where the issues/failures come from.

If the login with roundcubes succeded, please log-off again and switch back to Horde over the Plesk Panel. Normally Plesk will automatically change the webserver configs, but if this doesn't happen for unknown reasons, please rebuild the configs again with the above command with the httpdsmng - tool. Afterwards, please open again the URL http://webmail.YOURDOMAIN.COM and try to login to Horde now. If this isn't possible, again please investigate the error.logs, or provide entries from them, so that people willing to help may have a look at them. Be sure to have a look at the horde.log now, instead of the one from roundcube.

It is as well always helpfull to provide all informations what changes you did, to resolve problems/issues and always try to include your webserver configuration ( with or without FastCGI, with or without ngginx and so on... ), because some suggestions will fail, if used with a total different configuration and/or environment setup.

If you are not the thread starter and just post into the same thread, because you think you have the same problem/issue, be sure, that your own Plesk version, configuration and operating system meet the same environment like the thread starter, otherwise you should provide the different environment in your post in order to get help/suggestions/work-arounds for your problem/issue.

 

[Namesonic]

Plesk Panel 11.5.30 Update #47, last updated at Sept 28, 2014 11:24 PM
CentOS 6.5 (Final)

The autoinstaller completed successfully, and rebuilt webserver configs. Switched to roundcude. I tried in IE & FF -> webmail.domainname.com (Under plan A ) Got Bad Gateway/ 500 Internal Server Error after 45 seconds.

Adding -> FcgidMaxRequestLen 1073741824 -> restart -> no differences

nginx error log
2014/09/28 23:28:05 [error] 3103#0: *33 upstream prematurely closed connection while reading response header from upstream, client: 154.5.232.165, server: webmail.*, request: "GET /services/portal/index.php HTTP/1.1", upstream: "http://173.1.199.200:7080/services/portal/index.php", host: "webmail.xxxxxxx.com"
2014/09/28 23:28:14 [error] 3103#0: *33 recv() failed (104: Connection reset by peer) while reading response header from upstream, client: 154.5.232.165, server: webmail.*, request: "GET / HTTP/1.1", upstream: "http://n.n.n.n:7080/", host: "webmail.xxxxxxx.com"

apache error log
[Sun Sep 28 23:30:01 2014] [notice] Apache/2.2.15 (Unix) DAV/2 mod_fcgid/2.3.9 mod_jk/1.2.37 mod_python/3.3.1 Python/2.6.6 mod_ssl/2.2.15 OpenSSL/1.0.1e-fips mod_perl/2.0.4 Perl/v5.10.1 configured -- resuming normal operations
[Sun Sep 28 23:30:52 2014] [warn] [client n.n.n.n] mod_fcgid: read data timeout in 45 seconds
[Sun Sep 28 23:30:52 2014] [error] [client n.n.n.n] Premature end of script headers: index.php
[Sun Sep 28 23:31:04 2014] [warn] mod_fcgid: process 8251 graceful kill fail, sending SIGKILL
[Sun Sep 28 23:31:49 2014] [warn] [client n.n.n.n] mod_fcgid: read data timeout in 45 seconds
[Sun Sep 28 23:31:49 2014] [error] [client n.n.n.n] Premature end of script headers: index.php
[Sun Sep 28 23:31:58 2014] [warn] mod_fcgid: process 8262 graceful kill fail, sending SIGKILL

psa_horde.log
empty

/var/log/php-fpm/error.log
[28-Sep-2014 23:23:58] NOTICE: fpm is running, pid 29617
[28-Sep-2014 23:23:58] NOTICE: ready to handle connections

/var/log/plesk-roundcube/errors
empty

apache access_log
n.n.n.n - - [28/Sep/2014:23:46:24 -0700] "GET /services/portal/index.php HTTP/1.0" 500 608 "-" "Mozilla/5.0 (Windows NT 6.1; WOW64; rv:32.0) Gecko/20100101 Firefox/32.0"

==> /var/log/httpd/error_log <==
[Sun Sep 28 23:54:21 2014] [warn] [client n.n.n.n] mod_fcgid: read data timeout in 45 seconds
[Sun Sep 28 23:54:21 2014] [error] [client n.n.n.n] Premature end of script headers: index.php

==> /var/log/httpd/access_log <==
n.n.n.n - - [28/Sep/2014:23:53:36 -0700] "GET /services/portal/index.php HTTP/1.0" 500 608 "-" "Mozilla/5.0 (Windows NT 6.1; WOW64; rv:32.0) Gecko/20100101 Firefox/32.0"

==> /var/log/httpd/suexec_log <==
[2014-09-28 23:56:52]: uid: (502/horde_sysuser) gid: (506/506) cmd: cgi_wrapper

I ran the bootstrapper repair successfully. Webmail showed the login screen once, and went to a 500 internal server error page after 45 seconds. Now every attempt to browse to the webmail url gives a 500 internal server error page after 45 seconds.

What other logs are missing here? This isn't enough information to diagnose.

 

[UFHH01]

Hello Namesonic,

First I have to make clear, that "500" errors are not really specific, because they could result in a bunch of things. For these error - codes it is always necessary to investigate several logs, which then could point to specific problems/issues.

As you can see here:

[Sun Sep 28 23:30:52 2014] [warn] [client n.n.n.n] mod_fcgid: read data timeout in 45 seconds
and here:
[Sun Sep 28 23:54:21 2014] [warn] [client n.n.n.n] mod_fcgid: read data timeout in 45 seconds​

the requests were terminated, before the whole requested content could have been served ( Premature end of script headers: index.php ).

The standard configuration for "FcgidIOTimeout" in "fcgid.conf" could be raised to 90 or 120 seconds. This could avoid closing connections, struggling with several rewrites. To change the standard configuration for Centos, please edit "/etc/httpd/conf.d/fcgid.conf" and change the line "FcgidIOTimeout 45" to your needs. Please restart your apache2 after you changed the configuration ( /etc/init.d/httpd restart )

You could as well change the Plesk configurations for the domains in behalf of the 45 seconds limit, with the help of this KB - article: Domain shows 500 internal server error: mod_fcgid: read data timeout ( KB: 121251 )


Sometimes, your mySQL - server has some "bad" entries. This might result after migrations and misconfigurations, or unfinished updates/upgrades. The KB - article 121762 could help you, to delete these bad entries: Website unavailable error: mod_fcgid: read data timeout in 45 seconds ( KB: 121762 )


You could as well make sure, that your php5-fpm - socket for a specific (sub)domain is up and running. The command could be: ps -aux | grep fpm
If your specific (sub)domain is not listed, please reply again, so that people willing to help may suggest some work-arounds for this. You could as well have a look at your "last_httpd.conf" and "last_nginx.conf" to investigate if the problems/issues result of false entries for that (sub)domain.

 

[Milos_Vu]

Hello,

I'm having the same issues as mentioned above with horde only displaying log in screen and after entering creditials it times out. Tried everything from the posts, but nothing helps. This got broken after recent update because of bash security. Please help!



Mon Sep 29 18:20:48 2014] [warn] [client xxxxxxx] mod_fcgid: read data timeout in 45 seconds, referer: http://webmail.xxxxx.com/services/portal/ [Mon Sep 29 18:20:48 2014] [error] [client xxxxxxx] Premature end of script headers: index.php, referer: http://webmail.xxxxx.com/services/portal/ [Mon Sep 29 18:20:49 2014] [error] [client xxxxx] File does not exist: /usr/share/psa-horde/favicon.ico [Mon Sep 29 18:20:58 2014] [warn] mod_fcgid: process 7790 graceful kill fail, sending SIGKILL
xxxxxxx

 

[UFHH01]

Another possible reason might be caused because of wrong permissions for the wrapper or it's folders.

chown root:root /var/www/
chown root:root /var/www/cgi-bin/
chown root:root /var/www/cgi-bin/cgi_wrapper/cgi_wrapper*
chmod 755 /var/www/cgi-bin/cgi_wrapper

If this still doesn't help, please consider removing Horde and/or Roundcube completly and reinstall it again :

Debian/Ubuntu:
remove: sudo apt-get remove psa-passwd psa-ingo psa-mnemo psa-kronolith psa-turba psa-imp psa-horde plesk-roundcube
install: sudo apt-get install psa-passwd psa-ingo psa-mnemo psa-kronolith psa-turba psa-imp psa-horde plesk-roundcube

Redhat/Suse/Centos:
remove: rpm -e psa-passwd psa-ingo psa-mnemo psa-kronolith psa-turba psa-imp psa-horde plesk-roundcube
install: rpm -i psa-passwd psa-ingo psa-mnemo psa-kronolith psa-turba psa-imp psa-horde plesk-roundcube
​

or use the Plesk autoinstaller to do this:
remove: /usr/local/psa/admin/sbin/autoinstaller --select-product-id plesk --select-release-current --remove-component horde --remove-component roundcube
install: /usr/local/psa/admin/sbin/autoinstaller --select-product-id plesk --select-release-current --install-component horde --install-component roundcube --reinstall-patch​

I recommend always using the Plesk autoinstaller, because some components have depencies to other packages which the autoinstaller solves automatically, while with the apt-get / rpm solution you might sometimes have to modify your command line, depending on your previous system configurations.

 

[Milos_Vu]

Hello, after removing and installing horde and roundcube, I could login to the horde mail and use every option except mail. When i click on mail, I get the timeout after 10 seconds. It looks like its really broken?

 

[Alvarogar]

I have the same error. When i try to mail section the server responds with timeout

 

[UFHH01]

I know that it seems useless for some people, but please include errors from the logs, which you can find at different places on your server and don't forget to mention which operating system you use, as well as the Plesk version ( Parallels Plesk Panel for Linux services logs and configuration files ( KB: 111283 ) ), especially, when you changed something that has been suggested and still experience problems/issues, there might be another error now.

It is as well helpfull to include informations about your webserver setting ( apache 2.x , with or without nginx, with or without php5-fpm FastCGI ) so that people willing to help don't have to guess what sort of environment you are using. Some problems/issues don't necessarily result because of an update/upgrade, even that it's looking that way for you. Especially in threads with global errors like "500" it isn't clear, what the cause of the failure/problem/issue is, that's why investigating the logs is necessary and not mandatory.

 

[Milos_Vu]

Hello again, I have erased all my log files and when i try to view my webmail it doesnt populate with errors. Before the suggestions in the posts above, there were some time out .php scripts error, but now nothing it just doesn't work We are using the latest plesk panel 12 with all the updates on Cent OS 64 bit.

 

[UFHH01]

Hello Milos_Vu,

so you can open the URL http://webmail.YOURDOMAIN.COM and see the login page? Right?
And you are using apache2, nginx and php5-fpm ( FastCGI ) ?

What happens, when you insert the right login-name ( [email protected] ) and password? And what entries have you got in the error- and/or access - logs AFTER you tried to login?

Possible error- and access - logs for Centos with apache2, nginx, php5-fpm ( FastCGI ):

MySQL: /var/log/mysqld.log

Apache 2: /var/log/httpd/*log
NGINX: /var/log/nginx/*log
php5-fpm: /var/log/php5-fpm.log

Domain - Logs: /var/www/vhosts/system/<domain>/logs/*log

Courier-IMAP & Postfix & Qmail & Dovecot - Mail-Logs:
/usr/local/psa/var/log/maillog

Horde: /var/log/psa-horde/psa-horde.log

Roundcube: /var/log/plesk-roundcube/errors​

Are you aware, that IF you use fail2ban and if you turned on fail2ban - filters which monitor failed logins, scripts, apache - logs, nginx - logs, mail - logs and horde/roundcube - logs, that this may lead to a temporary ban of your IP while you experiment/configure your webserver and/or if you are testing things out which are misconfigured, so you may not open the URL http://webmail.YOURDOMAIN.COM and you may get the information, that the requested URL is not reachable? IF this is the case and you are still experimenting/configuring webserver stuff, please either whitelist your actual IP, or turn off fail2ban while you experiment/configure your webserver, to avoid being temporary banned.