Cannot start dnsadmin due to PID file issue.

[bobc02]

I'm using cPanel 88.0 (build 9).

The status of dnsadmin is DOWN. I am trying to start it, but getting PID file error. Here's results of a status check:

[root@dev scripts]# systemctl status dnsadmin
* dnsadmin.service - cPanel DNS admin service
   Loaded: loaded (/etc/systemd/system/dnsadmin.service; enabled; vendor preset: disabled)
   Active: failed (Result: protocol) since Fri 2020-06-12 18:18:34 UTC; 15s ago
  Process: 19634 ExecStart=/usr/local/cpanel/scripts/restartsrv_dnsadmin --no-verbose --notconfigured-ok --systemd-service=dnsadmin (code=exited, status=0/SUCCESS)

Jun 12 18:18:34 myserver.com systemd[1]: Starting cPanel DNS admin service...
Jun 12 18:18:34 myserver.com systemd[1]: Can't open PID file /var/run/dnsadmin.pid (yet?) after start: Too many levels of symbolic links
Jun 12 18:18:34 myserver.com systemd[1]: Failed to set a watch for dnsadmin.service's PID file /var/run/dnsadmin.pid: Too many open files
Jun 12 18:18:34 myserver.com systemd[1]: Failed to start cPanel DNS admin service.
Jun 12 18:18:34 myserver.com systemd[1]: Unit dnsadmin.service entered failed state.
Jun 12 18:18:34 myserver.com systemd[1]: dnsadmin.service failed.

I don't understand the "Too many levels of symbolic links", and "Too many open files", and how to correct the errors.

This seems to have started after an auto-update "PowerDNS Configuration updated", which immediately generated error "cPanel & WHM update failure in upcp script".

Any ideas on how to fix?

Thanks.

 

[andrew.n]

Looks like you are on OpenVZ virtualisation. This is cPanel advise for this problem:

Too many levels of symbolic links

CentOS 7.7 introduced an update to systemd that changed how systemd opens symbolic links. The update conflicts with out-of-date kernels on OpenVZ or Virtuozzo environments. Systems affected by this...

support.cpanel.net

The resolution could be either reaching out to your host to upgrade the kernel on OpenVZ server or as a workaround for example:

1. Edit the service configuration file:
   # systemctl edit --full spamassassin.service
2. Set the PIDfile parameter from /var/run/<service-name>.pid to /run/<service-name>.pid as follows:
   # systemctl cat spamassassin.service
   [Unit]
   Description=Spamassassin daemon
   After=syslog.target network.target
   
   [Service]
   Type=forking
   PIDFile=/run/spamd.pid
   EnvironmentFile=-/etc/sysconfig/spamassassin
   ExecStartPre=-/sbin/portrelease spamd
   ExecStart=/usr/bin/spamd --pidfile /run/spamd.pid $SPAMDOPTIONS
   StandardOutput=syslog
   StandardError=syslog
   Restart=always
   
   [Install]
   WantedBy=multi-user.target
3. Reload Systemctl:
   # systemctl daemon-reload
4. Start the service:
   # systemctl start spamassassin.service

 

[bobc02]

Thanks for the explanation and links!

I verified that OpenVZ was the issue, and contacted my provider who said that they are aware, and are planning to move all accounts to a new environment.

I have tried the workaround, but I am getting "Too many open files" when I do the reload.

Here's the file after changes:

[Unit]
Description=cPanel DNS admin service
After=cpanel.service
# dnsonly server

[Service]
Type=forking
# PIDFile=/var/run/dnsadmin.pid
PIDFile=/run/dnsadmin.pid
ExecStart=/usr/local/cpanel/scripts/restartsrv_dnsadmin --no-verbose --notconfigured-ok  --systemd-service=dnsadmin
ExecStop=/usr/local/cpanel/scripts/restartsrv_dnsadmin --stop --notconfigured-ok  --systemd-service=dnsadmin

[Install]
WantedBy=multi-user.target

And here are the actions I took:

[root@dev scripts]# systemctl edit --full dnsadmin.service
[root@dev scripts]# systemctl daemon-reload
[root@dev scripts]# systemctl start dnsadmin.service
Error: Too many open files
Job for dnsadmin.service failed. See "systemctl status dnsadmin.service" and "journalctl -xe" for details.

Thanks.
Bob

 

[RoseHosting]

Thanks for the explanation and links!

I verified that OpenVZ was the issue, and contacted my provider who said that they are aware, and are planning to move all accounts to a new environment.

I have tried the workaround, but I am getting "Too many open files" when I do the reload.

Here's the file after changes:

[Unit]
Description=cPanel DNS admin service
After=cpanel.service
# dnsonly server

[Service]
Type=forking
# PIDFile=/var/run/dnsadmin.pid
PIDFile=/run/dnsadmin.pid
ExecStart=/usr/local/cpanel/scripts/restartsrv_dnsadmin --no-verbose --notconfigured-ok  --systemd-service=dnsadmin
ExecStop=/usr/local/cpanel/scripts/restartsrv_dnsadmin --stop --notconfigured-ok  --systemd-service=dnsadmin

[Install]
WantedBy=multi-user.target

And here are the actions I took:

[root@dev scripts]# systemctl edit --full dnsadmin.service
[root@dev scripts]# systemctl daemon-reload
[root@dev scripts]# systemctl start dnsadmin.service
Error: Too many open files
Job for dnsadmin.service failed. See "systemctl status dnsadmin.service" and "journalctl -xe" for details.

Thanks.
Bob

You can append these lines to /etc/security/limits.conf file to increase the limit.

*       soft    nofile  35000
*       hard    nofile  65000

 

[bobc02]

Increasing the limits solved the problem. All services are now working.

Thanks Andrew and RoseHosting for all your help!

Bob

 

[andrew.n]

Glad to hear! Enjoy the weekend

 

[cPanelLauren]

Glad to see you were able to get the issue resolved and thanks to @andrew.n and @RoseHosting

 

[planetjoin]

You can append these lines to /etc/security/limits.conf file to increase the limit.

*       soft    nofile  35000
*       hard    nofile  65000

hello

I´ve tried all this and still get :
# systemctl start dnsadmin.service
Error: Too many open files
Job for dnsadmin.service failed. See "systemctl status dnsadmin.service" and "journalctl -xe" for details.

I´m using DNSONLY whm in virtuozzo enviroment and it now working properly.
First i got named errors with TCP_FASTOPEN :

Jul 11 11:24:26 ns2 named[19416]: automatic empty zone: view internal: 8.B.D.0.1.0.0.2.IP6.ARPA
Jul 11 11:24:26 ns2 named[19416]: automatic empty zone: view internal: EMPTY.AS112.ARPA
Jul 11 11:24:26 ns2 named[19416]: automatic empty zone: view internal: HOME.ARPA
Jul 11 11:24:26 ns2 named[19416]: none:104: 'max-cache-size 90%' - setting to 921MB (out of 1024MB)
Jul 11 11:24:26 ns2 named[19416]: set up managed keys zone for view external, file 'external.mkeys'
Jul 11 11:24:26 ns2 named[19416]: none:104: 'max-cache-size 90%' - setting to 921MB (out of 1024MB)
Jul 11 11:24:26 ns2 named[19416]: ../../../../lib/isc/unix/socket.c:5682: unexpected error:
Jul 11 11:24:26 ns2 named[19416]: setsockopt(23, TCP_FASTOPEN) failed with Protocol not available
Jul 11 11:24:26 ns2 named[19416]: command channel listening on 127.0.0.1#953
Jul 11 11:24:26 ns2 systemd[1]: Started Berkeley Internet Name Domain (DNS).

AND NOW.. dnsadmin can´t start.

Jul 11 11:40:19 ns2 systemd[1]: Starting cPanel DNS admin service...
Jul 11 11:40:19 ns2 systemd[1]: Can't open PID file /var/run/dnsadmin.pid (yet?) after start: Too many levels of symbolic links
Jul 11 11:40:19 ns2 systemd[1]: Failed to set a watch for dnsadmin.service's PID file /var/run/dnsadmin.pid: Too many open files
Jul 11 11:40:19 ns2 systemd[1]: Failed to start cPanel DNS admin service.
Jul 11 11:40:19 ns2 systemd[1]: Unit dnsadmin.service entered failed state.
Jul 11 11:40:19 ns2 systemd[1]: dnsadmin.service failed.

/var/run/dnsadmin.pid is not exists...



Any clue?

Thanks
Fabian

 

[andrew.n]

Open a support ticket with your provider and ask them to upgrade the kernel on the host your VPS resides.

 

[planetjoin]

Open a support ticket with your provider and ask them to upgrade the kernel on the host your VPS resides.

Thanks for your reply.

Finally , i fixed dnsadmin, but i still having issues with TCP_FASTOPEN. That is because this vps is a container with virtuozzo. The next week we will move it to a KVM vps to update kernels myself

Regards
Fabian