After installing boinc, I cannot keep boinc-client.service running.

Here is what happens:
Install boinc per:
http://boinc.berkeley.edu/wiki/Installing_BOINC_on_Fedora

using:

systemctl enable boinc-client.service

after booting:

sudo systemctl status boinc-client.service

returns:

boinc-client.service - Berkeley Open Infrastructure Network Computing Client
   Loaded: loaded (/usr/lib/systemd/system/boinc-client.service; enabled)
   Active: failed (Result: exit-code) since Mon 2013-05-27 07:27:31 CDT; 18min ago
  Process: 6838 ExecStopPost=/bin/rm -f /var/lib/boinc/lockfile (code=exited, status=0/SUCCESS)
  Process: 6815 ExecStart=/usr/bin/boinc --dir /var/lib/boinc (code=exited, status=76)
  Process: 6813 ExecStartPre=/usr/bin/chown boinc:boinc /var/log/boinc.log /var/log/boincerr.log (code=exited, status=0/SUCCESS)
  Process: 6811 ExecStartPre=/usr/bin/touch /var/log/boinc.log /var/log/boincerr.log (code=exited, status=0/SUCCESS)
  Process: 6809 ExecStartPre=/bin/sleep 1 (code=exited, status=0/SUCCESS)

May 27 07:27:01 localhost.localdomain systemd[1]: Started Berkeley Open Infrastructure Network Computing Client.
May 27 07:27:31 localhost.localdomain systemd[1]: boinc-client.service: main process exited, code=exited, status=76/n/a
May 27 07:27:31 localhost.localdomain systemd[1]: Unit boinc-client.service entered failed state.

then I issue:

sudo systemctl restart boinc-client.service
and:

sudo systemctl status boinc-client.service

and I get:

boinc-client.service - Berkeley Open Infrastructure Network Computing Client
   Loaded: loaded (/usr/lib/systemd/system/boinc-client.service; enabled)
   Active: active (running) since Mon 2013-05-27 07:47:31 CDT; 5s ago
  Process: 6838 ExecStopPost=/bin/rm -f /var/lib/boinc/lockfile (code=exited, status=0/SUCCESS)
  Process: 7606 ExecStartPre=/usr/bin/chown boinc:boinc /var/log/boinc.log /var/log/boincerr.log (code=exited, status=0/SUCCESS)
  Process: 7602 ExecStartPre=/usr/bin/touch /var/log/boinc.log /var/log/boincerr.log (code=exited, status=0/SUCCESS)
  Process: 7600 ExecStartPre=/bin/sleep 1 (code=exited, status=0/SUCCESS)
 Main PID: 7609 (boinc)
   CGroup: name=systemd:/system/boinc-client.service
           ├─7609 /bin/bash /usr/bin/boinc --dir /var/lib/boinc
           └─7610 /usr/bin/boinc_client --allow_multiple_clients --dir /var/lib/boinc

May 27 07:47:30 localhost.localdomain systemd[1]: Starting Berkeley Open Infrastructure Network Computing Client...
May 27 07:47:31 localhost.localdomain systemd[1]: Started Berkeley Open Infrastructure Network Computing Client.

and again:

sudo systemctl status boinc-client.service

returns:

boinc-client.service - Berkeley Open Infrastructure Network Computing Client
   Loaded: loaded (/usr/lib/systemd/system/boinc-client.service; enabled)
   Active: active (running) since Mon 2013-05-27 07:47:31 CDT; 16s ago
  Process: 6838 ExecStopPost=/bin/rm -f /var/lib/boinc/lockfile (code=exited, status=0/SUCCESS)
  Process: 7606 ExecStartPre=/usr/bin/chown boinc:boinc /var/log/boinc.log /var/log/boincerr.log (code=exited, status=0/SUCCESS)
  Process: 7602 ExecStartPre=/usr/bin/touch /var/log/boinc.log /var/log/boincerr.log (code=exited, status=0/SUCCESS)
  Process: 7600 ExecStartPre=/bin/sleep 1 (code=exited, status=0/SUCCESS)
 Main PID: 7609 (boinc)
   CGroup: name=systemd:/system/boinc-client.service
           ├─7609 /bin/bash /usr/bin/boinc --dir /var/lib/boinc
           └─7610 /usr/bin/boinc_client --allow_multiple_clients --dir /var/lib/boinc

May 27 07:47:30 localhost.localdomain systemd[1]: Starting Berkeley Open Infrastructure Network Computing Client...
May 27 07:47:31 localhost.localdomain systemd[1]: Started Berkeley Open Infrastructure Network Computing Client.

Now wait a bit:
and again issue:

sudo systemctl status boinc-client.service

and it returns:

boinc-client.service - Berkeley Open Infrastructure Network Computing Client
   Loaded: loaded (/usr/lib/systemd/system/boinc-client.service; enabled)
   Active: failed (Result: exit-code) since Mon 2013-05-27 07:48:02 CDT; 3min 35s ago
  Process: 7629 ExecStopPost=/bin/rm -f /var/lib/boinc/lockfile (code=exited, status=0/SUCCESS)
  Process: 7609 ExecStart=/usr/bin/boinc --dir /var/lib/boinc (code=exited, status=76)
  Process: 7606 ExecStartPre=/usr/bin/chown boinc:boinc /var/log/boinc.log /var/log/boincerr.log (code=exited, status=0/SUCCESS)
  Process: 7602 ExecStartPre=/usr/bin/touch /var/log/boinc.log /var/log/boincerr.log (code=exited, status=0/SUCCESS)
  Process: 7600 ExecStartPre=/bin/sleep 1 (code=exited, status=0/SUCCESS)

May 27 07:47:30 localhost.localdomain systemd[1]: Starting Berkeley Open Infrastructure Network Computing Client...
May 27 07:47:31 localhost.localdomain systemd[1]: Started Berkeley Open Infrastructure Network Computing Client.
May 27 07:48:02 localhost.localdomain systemd[1]: boinc-client.service: main process exited, code=exited, status=76/n/a
May 27 07:48:02 localhost.localdomain systemd[1]: Unit boinc-client.service entered failed state.

It fails after 30 second after the succesful restart!!

and /var/log/boincerr.log contains:

gstate.init() failed
Error Code: -180
gstate.init() failed
Error Code: -180
gstate.init() failed
.
.
.
.
.
Error Code: -180
gstate.init() failed
Error Code: -180

and all this before running boinmgr, and of course boincmgr does not run.

any insight?

Masoud.

I use F18, with all updates applied, on a X86_64 kernel with Evga Geforce 7600 GS PCI-E 256mb DDR2 video card.

-180 according the BFS

ERR_BIND -180

GUI RPC bind failed.

Is there no boinc-client script like there is for the Ubuntu RPM? The command on that platform is:

sudo service boinc-client start (or restart or stop)

Coincidently there's an other recent thread with a bind issue... the client already running

given the output parm --allow_multiple_clients, I'd figure that if port 31416 is already committed to one client, an additional would need another port. A wide swerve-ball guess.

---

Coelum Non Animum Mutant, Qui Trans Mare Currunt

-180 according the BFS

ERR_BIND -180

GUI RPC bind failed.

Is there no boinc-client script like there is for the Ubuntu RPM? The command on that platform is:

sudo service boinc-client start (or restart or stop)

Yes there is as I posted last and it is:

sudo systemctl restart boinc-client.service

This command, As I said restarts the service which stops within about 30 seconds!

Coincidentally there's an other recent thread with a bind issue... the client already running

Yes I did see that today. Our two posts might actually be the result of some update to the Fedora(?).
Because, I originally had the same problem two days before this post but an update to F18 seem to fix the problem. Then Boinc ran for that day, until another update to F18, the day of my post, may(?) have been the culprit to stop Boinc from working.

given the output parm --allow_multiple_clients, I'd figure that if port 31416 is already committed to one client, an additional would need another port. A wide swerve-ball guess.

Sorry, but I do not understand what you talking about:(

Thanks for you time.
Masoud

given the output parm --allow_multiple_clients, I'd figure that if port 31416 is already committed to one client, an additional would need another port. A wide swerve-ball guess.

Sorry, but I do not understand what you talking about:(

In your 3rd log in your first post, it says:

Main PID: 7609 (boinc)
CGroup: name=systemd:/system/boinc-client.service
├─7609 /bin/bash /usr/bin/boinc --dir /var/lib/boinc
└─7610 /usr/bin/boinc_client --allow_multiple_clients --dir /var/lib/boinc

See the --allow_multiple_clients attribute? That's what Rob was pointing out.

The client and the manager connect to each other through RPC on TCP port 31416. Only one client and one manager can use this port at the same time. As soon as you start a second client, it needs a separate port, e.g. 31417, and any manager connecting to it will need to use that port as well. Any further next client and manager combo will need a new port number.

When you run two or more clients, you can run into all kinds of weirdness. Including the closing of the client(s) for no apparent reason.

So check what is asking for the running of multiple clients and try to remove that attribute. See if things will run as is then. And then complain about this to the Fedora package maintainer.

---

given the output parm --allow_multiple_clients, I'd figure that if port 31416 is already committed to one client, an additional would need another port. A wide swerve-ball guess.

Sorry, but I do not understand what you talking about:(

In your 3rd log in your first post, it says:

Main PID: 7609 (boinc)
CGroup: name=systemd:/system/boinc-client.service
├─7609 /bin/bash /usr/bin/boinc --dir /var/lib/boinc
└─7610 /usr/bin/boinc_client --allow_multiple_clients --dir /var/lib/boinc

See the --allow_multiple_clients attribute? That's what Rob was pointing out.

The client and the manager connect to each other through RPC on TCP port 31416. Only one client and one manager can use this port at the same time. As soon as you start a second client, it needs a separate port, e.g. 31417, and any manager connecting to it will need to use that port as well. Any further next client and manager combo will need a new port number.

When you run two or more clients, you can run into all kinds of weirdness. Including the closing of the client(s) for no apparent reason.

So check what is asking for the running of multiple clients and try to remove that attribute. See if things will run as is then. And then complain about this to the Fedora package maintainer.

Fine, I see the point now.
Q: How do I "remove that attribute"?
I do not know how I have added that?

Sorry, if that is a FAQ.

Thanks to all for all the help.

---

Masoud.

I use F18, with all updates applied, on a X86_64 kernel with Evga Geforce 7600 GS PCI-E 256mb DDR2 video card.

I assume it'll be in the boinc-client.service script. If this is a script.

---

I assume it'll be in the boinc-client.service script. If this is a script.

I looked at that to no avail.

After that I rebooted the next day the service ran with no problem.

Logged off, to tend to something else. Came back rebooted, and could keep the service running trying every thing I could think of! Even though nothing was different form the previous time that it was running.

Thanks for all the help.

I will continue working on this, and post if I can find something.

---

Masoud.

I use F18, with all updates applied, on a X86_64 kernel with Evga Geforce 7600 GS PCI-E 256mb DDR2 video card.

If your question is: How do I get BOINC to start automatically after system reboot? do this:

systemctl enable boinc-client.service

This is how you cause an app to startup automatically after reboot.

---

www.usefulramblings.org

When you run two or more clients, you can run into all kinds of weirdness. Including the closing of the client(s) for no apparent reason.

It might be troublesome or impossible on Windows but I've done it easily and very successfully on Linux.

So check what is asking for the running of multiple clients and try to remove that attribute. See if things will run as is then. And then complain about this to the Fedora package maintainer.

The "allow multiple clients" param doesn't invoke multiple clients. It simply allows multiple clients to run if the user choses to do so. In other words you can use that param but start only one client. Indeed it makes little sense to use the param then start only one client but that's a different matter.