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

Re: [HTCondor-users] Need to "service condor restart" on the worker node to make server aware of worker node

On 21/11/17 17:38, Greg Thain wrote:
I'm moving to a Centos7 version of HTCondor, version 8.3.6.

I find I need to do "service condor restart" on the worker node to make server aware of worker node in the first place, else condor_status shows nothing.

Is that normal? Shouldn't service on worker node try to connect to the head node every now and then? Is there a time limit?

By default, the condor_startd updates the collector when it starts, every time there is a significant state change, and also every five minutes, in case the collector restarts. If the worker node is running fine, a restart shouldn't be needed.

Thanks, Greg. The issue was about the firewall on the systems - HTCondor needs to use UDP for the comms between STARTD and COLLECTOR. When I allowed that, things kicked into life.

BTW: The problem showed up in CollectorLog on Central Server as follows.

12/14/17 11:35:35 DC_AUTHENTICATE: attempt to open invalid session XXX:2246741:1513180756:9392, failing; this session was requested by <192.168.XXX:31506> with return address <192.168.XXX:9618?addrs=192.168.XXX-9618+[--1]-9618&noUDP&sock=32620_2f6d>


Steve Jones                             sjones@xxxxxxxxxxxxxxxx
Grid System Administrator               office: 220
High Energy Physics Division            tel (int): 43396
Oliver Lodge Laboratory                 tel (ext): +44 (0)151 794 3396
University of Liverpool                 http://www.liv.ac.uk/physics/hep/