502 Bad Gateway while click on Webspaces -> Add Webspace

[isp4p]

Hi,

Plesk 11.5.30 all latest updated installed is used with Power-User view on an Ubuntu 12.04 LTS 64 Bit system.


The Plesk Interface is working well in general. But if the customer clicks on "Webspaces" -> Add Webspace they run in a:

URL:
https://server:8443/smb/account/create-subscription

Browser:
502 Bad Gateway
nginx


To try to fix it with:

/usr/local/psa/admin/bin/nginxmng --disable
/usr/local/psa/admin/sbin/httpdmng --reconfigure-all
/etc/init.d/apache2 restart
/etc/init.d/ngnix restart ( does not work anyway since the customer is not using it ) -> ngnix_enable=no

Which did not work ( ngnix is and was not active )

In the log /var/log/sw-cp-server/error.log

we can see:

[error] 1936#0: *180 recv() failed (104: Connection reset by peer) while reading response header from upstream, client: 84.177.142.220, server: , request: "GET /smb/account/create-subscription HTTP/1.1", upstream: "fastcgi://unix:/var/run/sw-engine.sock:", host: "server:8443", referrer: "https://server:8443/smb/account"


Any idea how to fix that ?

Thank you and greetings
Oliver

 

[InderS]

Hello,

Please try to add following setting in your nginx.conf file

proxy_buffer_size 128k;
proxy_buffers 4 256k;
proxy_busy_buffers_size 256k;

 

[isp4p]

Hi,

thank you for your advice.

The ngnix deamon is not active:

/etc/init.d/nginx start
* Not starting NGINX as it is disabled in config

And its not supposed to be active, nor it was active before.

So i assume that changing in

/etc/ngnix/nginx.conf

will not do anything.


Or am i wrong ? Does the sw-cp-server require ngnix ?

Thank you !
Greetings
Oliver

 

[InderS]

Hello,

Can you please try this setting after enable the nginx on your server and let check if you are still facing the same issues in you Plesk panel

 

[isp4p]

hi,

/usr/local/psa/admin/bin/nginxmng --enable
/usr/local/psa/admin/sbin/httpdmng --reconfigure-all
/etc/init.d/apache2 restart
/etc/init.d/ngnix restart
/etc/init.d/sw-cp-server restart

proxy_buffer_size 128k;
proxy_buffers 4 256k;
proxy_busy_buffers_size 256k;

inserted.

Result:

502 Bad Gateway
nginx

Errorlog:

[error] 23472#0: *1 recv() failed (104: Connection reset by peer) while reading response header from upstream, client: 84.177.141.63, server: , request: "GET /smb/account/create-subscription HTTP/1.1", upstream: "fastcgi://unix:/var/run/sw-engine.sock:", host: "server:8443", referrer: "https://server:8443/smb/account"

Thank you !
Greetings
Oliver

 

[isp4p]

hi,

if it helps:

Switching from Poweruser to Service Provider overview gives for Adding clients / Domains:

[error] 23472#0: *14 recv() failed (104: Connection reset by peer) while reading response header from upstream, client: 84.177.141.63, server: , request: "GET /admin/subscription/create HTTP/1.1", upstream: "fastcgi://unix:/var/run/sw-engine.sock:", host: "server:8443", referrer: "https://server:8443/admin/domain/list?context=domains"

with 502 ngnix.

Greetings
Oliver

 

[DaRealDan]

hi,

if it helps:

Switching from Poweruser to Service Provider overview gives for Adding clients / Domains:

[error] 23472#0: *14 recv() failed (104: Connection reset by peer) while reading response header from upstream, client: 84.177.141.63, server: , request: "GET /admin/subscription/create HTTP/1.1", upstream: "fastcgi://unix:/var/run/sw-engine.sock:", host: "server:8443", referrer: "https://server:8443/admin/domain/list?context=domains"

with 502 ngnix.

Greetings
Oliver

I too have the exact same problem after upgrading from 11.0 to 11.5! Is there anything we can do?

 

[isp4p]

Hi,

yes, with us it happen also with an update.

The very last update: Panel Version: 11.5.30 Update Nr. 18 / OS: Ubuntu 12.04.3 LTS


Brings:

[error] 23472#0: *92 recv() failed (104: Connection reset by peer) while reading response header from upstream, client: 87.148.151.45, server: , request: "GET /smb/account/create-subscription HTTP/1.1", upstream: "fastcgi://unix:/var/run/sw-engine.sock:", host: "server:8443", referrer: "https://server:8443/smb/account"

After appport as application run several seconds on 100% CPU.

As it seems we will have to install a new server and migrate the accounts.

Its sad that it seems none of the officials cares about such problems.

Greetings
Oliver

 

[DaRealDan]

I have checked all possible log files

/opt/psa/admin/logs/httpsd_access_log

ip.add.re.ss - - [16/Oct/2013:10:00:01 +0200] "GET /admin/customer/create HTTP/1.1" 502 166 "https://server:8443/admin/customer/list?context=customers" "Mozilla/5.0 (Macintosh; Intel Mac OS X 10.7; rv:24.0) Gecko/20100101 Firefox/24.0" "-"'/admin/customer/create' '' '/opt/psa/admin/htdocs'

/var/log/sw-cp-server/error_log

2013/10/16 10:00:01 [error] 9384#0: *7 recv() failed (104: Connection reset by peer) while reading response header from upstream, client: ip.add.re.ss, server: , request: "GET /admin/customer/create HTTP/1.1", upstream: "fastcgi://unix:/var/run/sw-engine.sock:", host: "server:8443", referrer: "https://server:8443/admin/customer/list?context=customers"

After tweaking some error_logs I got the following from:

/var/log/sw-cp-server/sw-engine.log

[16-Oct-2013 10:17:16.157200] DEBUG: pid 9927, fpm_got_signal(), line 76: received SIGCHLD
[16-Oct-2013 10:17:16.157236] DEBUG: pid 9927, fpm_children_bury(), line 254: [pool plesk] child 9979 has been killed by the process managment after 60.347944 seconds from start
[16-Oct-2013 10:17:16.157247] DEBUG: pid 9927, fpm_event_loop(), line 419: event module triggered 1 events
[16-Oct-2013 10:17:17.151602] DEBUG: pid 9927, fpm_pctl_perform_idle_server_maintenance_heartbeat(), line 489: stopping idle maintenance as no children present ...
[16-Oct-2013 10:23:06.138008] DEBUG: pid 9927, fpm_got_signal(), line 85: received SIGTERM
[16-Oct-2013 10:23:06.138039] NOTICE: pid 9927, fpm_got_signal(), line 86: Terminating ...
[16-Oct-2013 10:23:06.138046] DEBUG: pid 9927, fpm_pctl(), line 232: switching to 'terminating' state
[16-Oct-2013 10:23:06.138050] NOTICE: pid 9927, fpm_pctl_exit(), line 71: exiting, bye-bye!
[16-Oct-2013 10:23:06.217486] DEBUG: pid 10060, fpm_scoreboard_init_main(), line 40: got clock tick '100'
[16-Oct-2013 10:23:06.217951] DEBUG: pid 10060, fpm_event_init_main(), line 337: event module is epoll and 105 fds have been reserved
[16-Oct-2013 10:23:06.217996] NOTICE: pid 10060, fpm_init(), line 83: fpm is running, pid 10060
[16-Oct-2013 10:23:06.218005] DEBUG: pid 10060, main(), line 1840: Sending "1" (OK) to parent via fd=11
[16-Oct-2013 10:23:06.218016] DEBUG: pid 10060, fpm_pctl_heartbeat(), line 461: heartbeat have been set up with a timeout of 200000ms
[16-Oct-2013 10:23:06.218020] DEBUG: pid 10060, fpm_pctl_perform_idle_server_maintenance_heartbeat(), line 489: stopping idle maintenance as no children present ...
[16-Oct-2013 10:23:06.218027] DEBUG: pid 10060, fpm_event_loop(), line 366: 61264 bytes have been reserved in SHM
[16-Oct-2013 10:23:06.218031] NOTICE: pid 10060, fpm_event_loop(), line 367: ready to handle connections
[16-Oct-2013 10:23:19.264234] DEBUG: pid 10060, fpm_children_make(), line 421: [pool plesk] child 10082 started
[16-Oct-2013 10:23:19.264275] DEBUG: pid 10060, fpm_pctl_on_socket_accept(), line 556: [pool plesk] got accept without idle child available .... I forked
[16-Oct-2013 10:23:19.264280] DEBUG: pid 10060, fpm_pctl_on_socket_accept(), line 572: starting idle maintenance...
[16-Oct-2013 10:23:19.264284] DEBUG: pid 10060, fpm_event_loop(), line 419: event module triggered 1 events
[16-Oct-2013 10:23:19.579971] WARNING: pid 10060, fpm_stdio_child_said(), line 167: [B][pool plesk] child 10082 said into stderr: "sw-engine-fpm: pool plesk: relocation error: /usr/lib/odbc/libmyodbc.so: symbol my_pthread_fastmutex_init, version libmysqlclient_16 not defined in file libmysqlclient_r.so.16 with link time reference"[/B]
[16-Oct-2013 10:23:19.580018] DEBUG: pid 10060, fpm_event_loop(), line 419: event module triggered 1 events
[16-Oct-2013 10:23:19.588299] DEBUG: pid 10060, fpm_event_loop(), line 419: event module triggered 1 events
[16-Oct-2013 10:23:19.588327] DEBUG: pid 10060, fpm_event_loop(), line 419: event module triggered 1 events
[16-Oct-2013 10:23:19.588955] DEBUG: pid 10060, fpm_got_signal(), line 76: received SIGCHLD
[16-Oct-2013 10:23:19.588980] WARNING: pid 10060, fpm_children_bury(), line 252: [pool plesk] child 10082 exited with code 127 after 0.324760 seconds from start
[16-Oct-2013 10:23:19.590138] NOTICE: pid 10060, fpm_children_make(), line 421: [pool plesk] child 10084 started
[16-Oct-2013 10:23:19.590149] DEBUG: pid 10060, fpm_event_loop(), line 419: event module triggered 1 events
[16-Oct-2013 10:23:20.265162] DEBUG: pid 10060, fpm_pctl_perform_idle_server_maintenance(), line 363: [pool plesk] currently 0 active children, 1 spare children

So there seems to be a problem with libmysqlclient16...

Any help?

 

[DaRealDan]

So I guess I found the problem. I have updated my Lucid 10.04 Ubuntu server with PHP 5.4.20 and MySQL 5.5 by using launchpad repositories which installed a new libmysqlclient18.

So after trying to install a new myodbc for my Ubuntu 10.04 (with possible support for libmysqlclient18 or so?) I gave up and tried one last thing... copying libmysqlclient_r.so.16 and libmysqlclient.so.16 from another server to the one with the problem. An guess what - it works. So the problem probably comes from libmysqlclient18?

 

[abdi]

Hi,

Plesk 11.5.30 all latest updated installed is used with Power-User view on an Ubuntu 12.04 LTS 64 Bit system.


The Plesk Interface is working well in general. But if the customer clicks on "Webspaces" -> Add Webspace they run in a:

URL:
https://server:8443/smb/account/create-subscription

Browser:
502 Bad Gateway
nginx


To try to fix it with:

/usr/local/psa/admin/bin/nginxmng --disable
/usr/local/psa/admin/sbin/httpdmng --reconfigure-all
/etc/init.d/apache2 restart
/etc/init.d/ngnix restart ( does not work anyway since the customer is not using it ) -> ngnix_enable=no

Which did not work ( ngnix is and was not active )

In the log /var/log/sw-cp-server/error.log

we can see:

[error] 1936#0: *180 recv() failed (104: Connection reset by peer) while reading response header from upstream, client: 84.177.142.220, server: , request: "GET /smb/account/create-subscription HTTP/1.1", upstream: "fastcgi://unix:/var/run/sw-engine.sock:", host: "server:8443", referrer: "https://server:8443/smb/account"


Any idea how to fix that ?

Thank you and greetings
Oliver

http://kb.parallels.com/en/118671

 

[abdi]

Hi,

Plesk 11.5.30 all latest updated installed is used with Power-User view on an Ubuntu 12.04 LTS 64 Bit system.


The Plesk Interface is working well in general. But if the customer clicks on "Webspaces" -> Add Webspace they run in a:

URL:
https://server:8443/smb/account/create-subscription

Browser:
502 Bad Gateway
nginx


To try to fix it with:

/usr/local/psa/admin/bin/nginxmng --disable
/usr/local/psa/admin/sbin/httpdmng --reconfigure-all
/etc/init.d/apache2 restart
/etc/init.d/ngnix restart ( does not work anyway since the customer is not using it ) -> ngnix_enable=no

Which did not work ( ngnix is and was not active )

In the log /var/log/sw-cp-server/error.log

we can see:

[error] 1936#0: *180 recv() failed (104: Connection reset by peer) while reading response header from upstream, client: 84.177.142.220, server: , request: "GET /smb/account/create-subscription HTTP/1.1", upstream: "fastcgi://unix:/var/run/sw-engine.sock:", host: "server:8443", referrer: "https://server:8443/smb/account"


Any idea how to fix that ?

Thank you and greetings
Oliver

http://kb.parallels.com/en/118671

 

[SoftCreatR]

I'm currently facing exactly the same sympthoms. However, i don't think, that they're related to the MySQL thing, since i havent found any error in sw-engine.log.