Mailing List Archives

Public Access

 		[Computer Systems Lab]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [HTCondor-users] condor start after installation on CentOS 7

I have used sudo condor_master and then the condor status changed.

[labounek@alela8 ~]$ condor_status
Name               OpSys      Arch   State     Activity LoadAv Mem   ActvtyTime

slot1@alela8. LINUX      X86_64 Unclaimed Idle      0.000  977  0+00:04:41
slot2@alela8. LINUX      X86_64 Unclaimed Idle      0.000  977  0+00:05:05
slot3@alela8. LINUX      X86_64 Unclaimed Idle      0.000  977  0+00:05:06
slot4@xxxxxxxx LINUX      X86_64 Unclaimed Idle      0.000  977  0+00:05:07
slot5@alela8. LINUX      X86_64 Unclaimed Idle      0.000  977  0+00:05:08
slot6@alela8. LINUX      X86_64 Unclaimed Idle      0.000  977  0+00:05:09
slot7@alela8. LINUX      X86_64 Unclaimed Idle      0.000  977  0+00:05:10
slot8@alela8. LINUX      X86_64 Unclaimed Idle      0.000  977  0+00:05:03
                     Machines Owner Claimed Unclaimed Matched Preempting

        X86_64/LINUX        8     0       0         8       0          0

               Total        8     0       0         8       0          0
[labounek@alela8 ~]$

The output of the next command also changed and it seems to look good.

[labounek@alela8 ~]$ ps -ef | grep condor
condor   24159     1  0 21:20 ?        00:00:00 condor_master
root     24160 24159  0 21:20 ?        00:00:00 condor_procd -A /var/run/condor/procd_pipe -L /var/log/condor/ProcLog -R 1000000 -S 60 -C 992
condor   24161 24159  0 21:20 ?        00:00:00 condor_collector -f
condor   24162 24159  0 21:20 ?        00:00:00 condor_negotiator -f
condor   24163 24159  0 21:20 ?        00:00:00 condor_schedd -f
condor   24164 24159  0 21:20 ?        00:00:00 condor_startd -f
labounek 24370 23779  0 21:28 pts/0    00:00:00 grep --color=auto condor
[labounek@alela8 ~]$

The masterlog also seems good.

12/13/14 21:20:07 ******************************************************
12/13/14 21:20:07 ** condor_master (CONDOR_MASTER) STARTING UP
12/13/14 21:20:07 ** /usr/sbin/condor_master
12/13/14 21:20:07 ** SubsystemInfo: name=MASTER type=MASTER(2) class=DAEMON(1)
12/13/14 21:20:07 ** Configuration: subsystem:MASTER local:<NONE> class:DAEMON
12/13/14 21:20:07 ** $CondorVersion: 8.2.5 Dec 01 2014 BuildID: UW_development $
12/13/14 21:20:07 ** $CondorPlatform: X86_64-RedHat_7.0 $
12/13/14 21:20:07 ** PID = 24159
12/13/14 21:20:07 ** Log last touched 12/11 16:35:39
12/13/14 21:20:07 ******************************************************
12/13/14 21:20:07 Using config source: /etc/condor/condor_config
12/13/14 21:20:07 Using local config sources:
12/13/14 21:20:07    /etc/condor/condor_config.local
12/13/14 21:20:07 config Macros = 58, Sorted = 58, StringBytes = 1461, TablesBytes = 2128
12/13/14 21:20:07 CLASSAD_CACHING is OFF
12/13/14 21:20:07 Daemon Log is logging: D_ALWAYS D_ERROR
12/13/14 21:20:07 DaemonCore: command socket at <147.229.74.20:48624>
12/13/14 21:20:07 DaemonCore: private command socket at <147.229.74.20:48624>
12/13/14 21:20:07 Master restart (GRACEFUL) is watching /usr/sbin/condor_master (mtime:1417472014)
12/13/14 21:20:07 Started DaemonCore process "/usr/sbin/condor_collector", pid and pgroup = 24161
12/13/14 21:20:07 Waiting for /var/log/condor/.collector_address to appear.
12/13/14 21:20:08 Found /var/log/condor/.collector_address.
12/13/14 21:20:08 Started DaemonCore process "/usr/sbin/condor_negotiator", pid and pgroup = 24162
12/13/14 21:20:08 Started DaemonCore process "/usr/sbin/condor_schedd", pid and pgroup = 24163
12/13/14 21:20:08 Started DaemonCore process "/usr/sbin/condor_startd", pid and pgroup = 24164

Last question, is it possible to start sudo condor_master automatically after reboot? I am not truly sure if does sudo commands saved in file /etc/rc.d/rc.local work or not. Or does another way exist?

Thanks,
Rene Labounek



OLD MESSAGES
_________________________________________________________________________________________________________________

Dne 11.12.2014 17:30, Ben Cotton napsal(a):
On Thu, Dec 11, 2014 at 11:09 AM, René Labounek
<xlabou01@xxxxxxxxxxxxxxxxxx> wrote:

Does anyone have some idea what could be wrong?

Does the master log (`condor_config_val MASTER_LOG` for location) have
any output? Since you installed via yum and didn't do any optional
configuration, I wouldn't expect missing libraries or invalid
configuration to be a problem, but that might be the case. If the
master log is empty, try running condor_master  from the command line.


Thanks,
BC
MASTER_LOG variable exits.

[labounek@alela8 ~]$ condor_config_val MASTER_LOG
/var/log/condor/MasterLog
[labounek@alela8 ~]$ nano /var/log/condor/MasterLog

Masterlog file is rewritten each 10 seconds with following message which ends with some error. It hase over 46000 rows now with still same text only different date and time, can I delete the whole text inside?

12/11/14 12:42:40 ******************************************************
12/11/14 12:42:40 ** condor_master (CONDOR_MASTER) STARTING UP
12/11/14 12:42:40 ** /usr/sbin/condor_master
12/11/14 12:42:40 ** SubsystemInfo: name=MASTER type=MASTER(2) class=DAEMON(1)
12/11/14 12:42:40 ** Configuration: subsystem:MASTER local:<NONE> class:DAEMON
12/11/14 12:42:40 ** $CondorVersion: 8.2.5 Dec 01 2014 BuildID: UW_development $
12/11/14 12:42:40 ** $CondorPlatform: X86_64-RedHat_7.0 $
12/11/14 12:42:40 ** PID = 13986
12/11/14 12:42:40 ** Log last touched 12/11 12:42:35
12/11/14 12:42:40 ******************************************************
12/11/14 12:42:40 Using config source: /etc/condor/condor_config
12/11/14 12:42:40 Using local config sources:
12/11/14 12:42:40    /etc/condor/condor_config.local
12/11/14 12:42:40 config Macros = 58, Sorted = 58, StringBytes = 1457, TablesBytes = 2128
12/11/14 12:42:40 CLASSAD_CACHING is OFF
12/11/14 12:42:40 Daemon Log is logging: D_ALWAYS D_ERROR
12/11/14 12:42:40 ERROR "can't safe_open_wrapper(/var/lock/condor/InstanceLock,O_WRONLY|O_CREAT|O_APPEND,S_IRUSR|S$
12/11/14 12:42:45 ******************************************************

I have found similar problem on MAC OS here: http://twiki.cis.rit.edu/twiki/bin/view/Help/CondorOnMacOS

CONDOR_CONFIG variable does not exist on my computer.

[labounek@alela8 ~]$ condor_config_val CONDOR_CONFIG
Not defined: CONDOR_CONFIG

Is it right that CONDOR_CONFIG variable should be set to /etc/condor/condor_config? Is next syntax correct?
condor_config_val  -set "CONDOR_CONFIG = /etc/condor/condor_config"

In condor_config file is then defined variable LOCAL_CONFIG_FILE = /etc/condor/condor_config.local where I think I can define properties unique for a given computer in grid. For example, variable describing if a given computer has or has not a cuda graphic card. Just now, file /etc/condor/condor_config.local does not exist on my computer.

When I wrote next comand I have got a message that files are setted properly:

[labounek@alela8 condor]$ condor_config_val -config
Configuration source:
        /etc/condor/condor_config
Local configuration source:
        /etc/condor/condor_config.local

Thanks,
RL


Dne 11.12.2014 17:25, Rich Pieri napsal(a):
On 12/11/14 11:09 AM, René Labounek wrote:

Does anyone have some idea what could be wrong?

Debugging issues with systemd can be troublesome because it doesn't use
conventional log files. Instead, it uses something called the systemd
journal. You'll need to install journalctl if it isn't already installed
and use that to examine the journal.

Good luck.

Terminal wrote to me that package journalctl is not available. But I think that problem is there where Ben Cotton was writing.