Hi, Am 11.08.2014 um 16:16 schrieb Kerim Gueney:
> Hello guys, > > Absolute gridengine newbie here. I'm trying to fix a problem that we have > with our gridengine, while our main admin is on his vacation. We are using > Debian stable running gridengine-master (6.2u5-7.1). The Debian supplied packages are known to cause some issues. The usual suggestion is to use a the SoGE source and compile on your own or use one of its packages of SoGE. > The problem is that gridengine-master won't start, at all. if I call it > manually it returns nothing. Setting the SGE_ND environment variable results > in the following output (blank error): > > # /etc/init.d/gridengine-master start > error: > > # > > > qstat shows > > # qstat > error: unable to read qmaster name: qmaster hostname in > "/var/lib/gridengine/default/common/act_qmaster" has zero length During startup the "act_qmaster" file is first read and then possibly adjusted and filled with the name of the machine the startup was done on, mainly to support the case of a failover setup with two or more masters. > adding the hosts name to the file manually doesn't help. It results in > > # qstat > error: commlib error: got select error (Connection refused) > error: unable to send message to qmaster using port 6444 on host > "queuemaster": got send error This name `queuemaster` is the name of the machine also when it's looked up by `hostname` or in /etc/hosts? In /usr/sge/default/common/bootstrap is "ignore_fqdn true" set (resp. you location of the file) (don't change it if it's not the case!)? I assume this, as "queuemaster" above has no domain attached. > attached is an strace of output of > > # strace /etc/init.d/gridengine-master start I don't know what all these calls to Debian procedures are doing. Hence let's try a different investigation: 1) is the name resolution working in a proper way? I.e. `hostname` gives the correct information; is it in short or FQDN by default? 2) Is there any file in /tmp of the "qmasterd" and attached PID with any hint? -- Reuti > I'm grateful for every kind of help I can get. Thank you in advance. > > Best regards, > Kerim Gueney > <strace_gridengine.txt>_______________________________________________ > users mailing list > [email protected] > https://gridengine.org/mailman/listinfo/users _______________________________________________ users mailing list [email protected] https://gridengine.org/mailman/listinfo/users
