[torqueusers] log file explosion

Martin Siegert siegert at sfu.ca
Tue Oct 27 14:15:10 MDT 2009


Hi Ken,

that's the strange thing: the log_level is set to 0

Thus, you are telling me that I shouldn't see these messages
at all, right?

# qmgr -c 'p s' | grep log
set server log_events = 511
set server log_level = 0

Cheers,
Martin

On Tue, Oct 27, 2009 at 01:47:30PM -0600, Ken Nielson wrote:
> Martin,
>
> What is your log level set to? I have looked at these messages and they 
> range from 2 to 6. What level of logging do you need?
>
> Ken Nielson
> Adaptive Computing
>
>
> Martin Siegert wrote:
>> Hi,
>>
>> ever since we upgraded torque on one of our clusters to 2.3.7
>> the log file size (in /var/spool/torque/server_logs) exploded
>> to about 50 times its former size.
>>
>> The logs are full of messages like these:
>>
>> 10/26/2009 00:00:03;0040;PBS_Server;Req;do_rpp;rpp request received on stream 9
>> 10/26/2009 00:00:03;0040;PBS_Server;Req;do_rpp;inter-server request received
>> 10/26/2009 00:00:03;0004;PBS_Server;Svr;is_request;message received from stream
>> 9 (version 1)
>> 10/26/2009 00:00:03;0004;PBS_Server;Svr;is_request;message received from stream
>> 172.21.0.11:15003
>> 10/26/2009 00:00:03;0004;PBS_Server;Svr;is_request;message STATUS (4) received f
>> rom mom on host r11 (172.21.0.11:15003) (stream 9)
>> 10/26/2009 00:00:03;0004;PBS_Server;Svr;is_request;IS_STATUS received from r11
>> 10/26/2009 00:00:03;0040;PBS_Server;Req;is_stat_get;received status from node r1
>> 1
>> 10/26/2009 00:00:03;0040;PBS_Server;Req;update_node_state;adjusting state for no
>> de r11 - state=0, newstate=0
>> 10/26/2009 00:00:03;0004;PBS_Server;Svr;is_request;node 'r11' is at state '0x0'
>>
>> This repeats itself for each node every second or so.
>> Ther settings are:
>> # qmgr -c 'p s' | grep log
>> set server log_events = 511
>> set server log_level = 0
>>
>> However on a different cluster also running 2.3.7 no such log entries
>> exist - the log contains information about jobs only. On that server:
>> # qmgr -c 'p s' | grep log
>> set server log_events = 511
>>
>> i.e., the log_level is not set.
>> I do not know whether this is relevant or not, but I cannot unset the
>> log_level:
>>
>> # qmgr -c 'unset server log_level'
>> # qmgr -c 'p s' | grep log
>> set server log_events = 511
>> set server log_level = 0
>>
>> Does anybody know what is causing these log entries and how I can get
>> rid of them (other than reverting back to 2.1.x)?
>>
>> Cheers,
>> Martin


More information about the torqueusers mailing list