[GE users] Multiple problems with CSP installation

Prentice Bisbal prentice at ias.edu
Fri Dec 5 16:23:23 GMT 2008


I'm attempting to do an automated installation of GE 6.2 with CSP
enabled. A previous installation a couple of days ago worked fine, and
now I'm just trying to duplicate that installation with CSP enabled. I
have encountered multiple problems. I installed a 5.3 system with CSP a
few years ago w/o a single problem, but I also wasn't doing automated
installs (very small cluster).

Is CSP broken in 6.2?

When I install the master, sge_execd won't start after an automatic
install like this

cd $SGE_ROOT
./inst_sge -csp -m -auto /path/to/file

The log file shows an error like this:

<quote>

error: commlib error: can't set CA chain file
(/var/sgeCA/port6444/default/userkeys/root/cert.pem)
error: commlib error: ssl error ([ID=33558530] in module "system
library": "No such file or directory")
unable to send message to qmaster using port 6444 on host
"aurora.sns.ias.edu": can't set CA chain file

Command failed: ./bin/lx24-amd64/qconf -Ahgrp /tmp/hostqueue1215

Probably a permission problem. Please check file access permissions.
Check root read/write permission. Check if SGE daemons are running.

</quote>

I checked in /var/sgeCA, and found that the directory
/var/sgeCA/port6444 doesn't even exist. Shouldn't it be created
automatically?

When do I do manual (interactive) install like this:

cd SGE_ROOT
./install_qmaster -csp

at the step where sge_execd starts, it takes a long time to start. In
fact, I get an error that it didn't start, but it did eventually start,
because I see sge_qmaster in the output of 'ps -ef'. Afterwards, i can
stop/start sge_qmaster w/o a problem. This doesn't happen during a
non-csp install.

I also have problems automatically installing the execd nodes
automatically. Again, this worked fine without the -csp option. The log
file doesn't show anything useful.

Am I doing something completely wrong? I did not copy /var/sgeCA over to
all the exec hosts before running the install, but it's my understanding
that the automatic install does this for me using scp.

-- 
Prentice

------------------------------------------------------
http://gridengine.sunsource.net/ds/viewMessage.do?dsForumId=38&dsMessageId=91397

To unsubscribe from this discussion, e-mail: [users-unsubscribe at gridengine.sunsource.net].



More information about the gridengine-users mailing list