Change server_host in Qmgr

Users/Site Administrators
1

Help!
I’m using Ansys Fluent 21.2 on a cluster and use PBS.
Since we need to move to Ubuntu (from CentOS), I need to recompile. Everything works fine, but…
Connection issues.
When I was checking the configs, I found the following odd thing:
Running qmgr -c “list server”:

Server tst1sjbox
server_state = Active
server_host = ubuntuheadv3.internal.cloudapp.net
scheduling = True
total_jobs = 20
state_count = Transit:0 Queued:0 Held:0 Waiting:0 Running:0 Exiting:0 Begun:0
managers = @*
operators = n@*
default_queue = workq
log_events = 511
mailer = /usr/sbin/sendmail
mail_from = adm
query_other_jobs = True
resources_default.ncpus = 1
default_chunk.ncpus = 1
scheduler_iteration = 600
flatuid = True
resv_enable = True
node_fail_requeue = 310
max_array_size = 10000
pbs_license_min = 0
pbs_license_max = 2147483647
pbs_license_linger_time = 31536000
license_count = Avail_Global:1000000 Avail_Local:1000000 Used:0 High_Use:0
pbs_version = 20.0.0
eligible_time_enable = False
job_history_enable = True
max_concurrent_provision = 5
power_provisioning = False
max_job_sequence_id = 9999999

The problem: my server_host is invalid. It should be tst1sjbox.internal.cloudapp.net
(ubuntuheadv3 is the original name of the image creation.)
How can I change this server_host entry?

I tried “set server tst1sjbox server_host=tst1sjbox.internal.cloudapp.net” but that just returns, that I don’t have enough privileges (while I’m running qmgr as root).

When I start a job from Fluent (via Windows GUI) the jobs are started on the nodes, but the connection times out. (MPT Connection Timeout).

2

You cannot change this value via qmgr.
what is the output of hostname command on the PBS Server host ?
Also please share the /etc/hosts contents against he PBS Server host

Even with privileges this cannot be changed using qmgr

3

Hi,

The /etc/hosts file is default (no entries in there but 127.0.0.1 localhost)
I moved the datastore to a backup location and reinstalled the PBS. That worked. I suppose the post-install script is taking care of the postgresql entry server_host.
I only have to solve the mpt_connect timeout issue.

4

Please make sure PBS Server host and Compute node hosts have a static IP and hostname.
Do not use the loopback address . Also, please make sure that these static IPs and hostnames are part of /etc/hosts file and are the same across server and compute nodes.

5

Well, the hosts file is equal on all head/ and compute nodes and contain the static IP’s and hostnames.
The Jobs are running. Qstat reports running time and the processes are visible on the compute nodes using 100% cpu.
My guess that there’s something wrong with the reporting from PBS to the client (ANSYS).
No firewall is installed but it looks like it’s a one-way direction.
But I will start/use another topic for that.
Appearantly the server_hostname cannot be changed during runtime easily.

Thanks for the help.