Plesk 12.0.18 MU#63

[tkalfaoglu]

Hi tkalfaoglu,


Could you please post as well your output of:

grep ^processor /proc/cpuinfo | wc -l​

Yes, it has 8 cores..

 

[IgorG]

Please tell me your OS version, output of following commands depending of your OS:

# rpm -qa | grep proftpd
# dpkg -l | grep proftpd

and attach your /tmp/autoinstaller3.log here.

 

[greos]

Hi IgorG,

Firstly, trying last ...6 hours to make it work again I did an update at proftpd via console (yum upgrade psa-proftpd)... no success.
This is what I have now:

OS: Centos 6.5 (final)

[root@origin ~]# rpm -qa | grep proftpd
psa-proftpd-1.3.6-8.el6.art.x86_64
psa-proftpd-xinetd-1.3.6-8.el6.art.x86_64

Please check autoinstaller file

 

[IgorG]

You have packages from Atomic repository but not packages released by Parallels for proftpd 1.3.5a version.
Contact Atomic regarding your issue.

 

[greos]

The problem started before did the update via console today, that probably installed via Atomic repository.
That was also the reason that I did the update to 12.5.30 today, in case that this will fix the problem who started
by update #63 on 12.0.18 two days ago.

 

[UFHH01]

Hi tkalfaoglu,

to optimize nginx, it is wise to use as many "worker_processes" as CPU cores and to raise the standard "worker_connections" from 768 to something more suitable, depending to your hosted sites on the server. As far that I remember, you have more than 300 sites running on your server, so please increase "worker_connections" to for example "10240"

events {

    worker_connections  10240;
    }

... and consider raising as well the "worker_rlimit_nofile" to about 64000.

It is wise to set the open files limit for nginx directly in your system configuration "/etc/security/limits.conf" and please make sure, that there is no standard restriction at for example in "90-nproc.conf", when you use CentOS ( standard there is: "* soft nproc 1024", where "*", is used for ALL users on the system ). You can define user - specific settings as for example to nginx, www-data ( "apache" for CentOS ), mysql. A decent setting is for example "64000", when you have got 8 CPU cores.

 

[IgorG]

Guys, regarding nginx start issue - it is really bug (PPPM-3259 for your reference).
Workaround: just add NGINX_ENABLED=yes to /etc/sysconfig/nginx file.

 

[tkalfaoglu]

Many thanks, that was great help.. Updating configs now..

 

[Kyobul]

For people still dealing with the problem, looking at /var/log/messages will tell you exactly where it bugs. I've commented that include line and everything is ok.

On my version: rpm -qa | grep proftpd shows :
psa-proftpd-xinetd-1.3.6-8.el6.art.x86_64
psa-proftpd-1.3.6-8.el6.art.x86_64

 

[Stephen_Stephen]

same problem from 5th Sep, error from /var/log/messages:

/etc/proftpd.d/ no such directory error in #95 /etc/proftpd.conf

system:
CentOS 6.7
12.0.18 Update #62

# rpm -qa | grep proftpd
psa-proftpd-1.3.6-8.el6.art.x86_64
psa-proftpd-xinetd-1.3.6-8.el6.art.x86_64

# rpm -q --changelog psa-proftpd
Fri Sep 04 2015 Support
- upgrade to 1.3.5a

 

[IgorG]

Try to disable Atomic repo at least and install proftpd packages from Odin repository.

 

[greos]

Currently from my side, I've just re-install the whole server (it's brand new).

Igor, just for your information, the update procedure was via Plesk, without touch anything on the files or change repository before or during the update.
Between 2nd to 5th of Sep. there was two updates on 12.0.18 regarding proftpd.
The first update did the ...damage. The second, didn't solve it.

 

[Stephen_Stephen]

Just tried commenting out the line include the /etc/proftpd.d/ it works now. thanks for kyobul

 

[Dx3webs]

Just a heads up.. we are seeing this message on various servers

proftpd[8023]: error: cannot read configuration path '/etc/proftpd.d': Not a directory
proftpd[8023]: fatal: Include: error including '/etc/proftpd.d/*.conf': Invalid argument on line 95 of '/etc/proftpd.conf'

possibly a problem with updates?

Creating /etc/proftpd.conf manually and restarting xinetd is a quick fix

 

[justausername]

Hello, This update has broken FTP on our server. I tried commenting out the "/etc/proftpd.d/*.conf" line in our /etc/proftpd.conf and restarting xinetd but when I try to connect with FTP it just times out while waiting for the directory listing. Not sure what to do next.

# rpm -qa | grep proftpd
psa-proftpd-1.3.5a-cos6.build1200150901.11.x86_64
# dpkg -l | grep proftpd
-bash: dpkg: command not found

OSCentOS 6.5 (Final)
Plesk version12.0.18 Update #63, last updated at Sept 5, 2015 03:35 AM​

 

[greos]

Just for update, seems that everything running fine now regarding proftpf.
I was lucky enough as the server is brand new, prepared for migration, so a clean install (3rd time) to 12.0.18 and then upgrade to 12.5.30 was an easy story.
Proftpd works as expected now, but really I'm curious for the reason that makes proftpd to reject connections after the #62 or #63 update.
At least, I'm not the only one

(Comment out the Include line didn't work at my case)

 

[justausername]

OK, I figured out my issue. The Plesk update overwrote the part of my proftpd.conf that included my passive ports. Once I put that back in everything is working fine for me again.

 

[ESTUGO]

same problem with "disabled in config", but only in CentOS 6.6/6.7, CentOS 7 not affected- my workaround for CentOS 6:

echo 'NGINX_ULIMIT="-n 6144"' >> /etc/sysconfig/nginx && /usr/local/psa/admin/sbin/httpdmng --reconfigure-all && /usr/local/psa/admin/bin/nginxmng --enable

I have a new problem, not all servers will be updated to nginx 1.9, here a sample:

[root@www nginx]# yum list installed | grep ngin
sw-engine.x86_64 2.13.12-centos6.201502091549 @PSA_12_0_18-dist
sw-nginx.x86_64 1.6.0-1.14051516.centos6 @NGINX_1_6_0-dist

[root@www ~]# cat /usr/local/psa/version
12.0.18 CentOS 6 1200150817.13

[root@www ~]# cat /root/.autoinstaller/microupdates.xml
<?xml version="1.0" encoding="UTF-8" standalone="yes" ?>
<patches>
<product id="plesk" version="12.0.18" installed-at="20150420T091830">
<patch version="63" timestamp="" installed-at="20150904T175725" />
</product>
</patches>

[root@www ~]# /usr/local/psa/admin/sbin/autoinstaller --select-release-current --upgrade-installed-components
You already have the latest version of product(s) and all selected components
installed. Installation will not continue.
Doing restart of Parallels Installer...

###############

This servers hase the same OS-version, but nginx is 1.9

[root@www ~]# cat /usr/local/psa/version
12.0.18 CentOS 6 1200150817.13

[root@www38 ~]# cat /root/.autoinstaller/microupdates.xml
<?xml version="1.0" encoding="UTF-8" standalone="yes" ?>
<patches>
<product id="plesk" version="12.0.18" installed-at="20150428T094417">
<patch version="63" timestamp="" installed-at="20150904T055159" />
</product>
</patches>

[root@www38 ~]# yum list installed | grep ngin
sw-engine.x86_64 2.13.12-centos6.201502091549 @PSA_12_0_18-dist
sw-nginx.x86_64 1.9.4-centos6.15090210 @NGINX_1_9_4-dist


################

Why can not upgrade nginx, all updates are available? Same OS and Plesk version

thanks for your help!

 

[IgorG]

Why can not upgrade nginx, all updates are available? Same OS?

Try to do following:

# /usr/local/psa/admin/sbin/autoinstaller --select-release-current --remove-component nginx
# /usr/local/psa/admin/sbin/autoinstaller --select-release-current --install-component nginx

and check result.

 

[ESTUGO]

Hi Igor,

commands executed:

# /usr/local/psa/admin/sbin/autoinstaller --select-release-current --remove-component nginx
# /usr/local/psa/admin/sbin/autoinstaller --select-release-current --install-component nginx

Patches were installed successfully.
The changes were applied successfully.

yum list installed | grep ngin
sw-engine.x86_64 2.13.12-centos6.201502091549 @PSA_12_0_18-dist
sw-nginx.x86_64 1.6.0-1.14051516.centos6 @NGINX_1_6_0-dist

same problem . In the CLI output I have seen downloading 1_6_0 from plesk-repo. Any idea?