[orbit-user] Time Issues on the Grid?

[Ivan Seskar]

Hi Rick,

Grid DHCP servers are advertising the console as a time reference so
that nodes can use it and I thought that all of our baseline images rely
on it. Will look into it a bit more tomorrow.

Ivan. 

-----Original Message-----
From: orbit-user-bounces at orbit-lab.org
[mailto:orbit-user-bounces at orbit-lab.org] On Behalf Of Rick Correa
Sent: Sunday, June 08, 2008 4:41 PM
To: Orbit user discussion mailing list
Subject: [orbit-user] Time Issues on the Grid?

I received the following error on the grid:

ERROR NodeApp: tar: proj/linkMonitorETX.py: time stamp 2008-06-07
18:24:39 is 108074063.361408532 s in the future

for each file in my tarball.  Is this an error or a warning?  

I'm guessing there is some time sync issues on the grid.  I saw on some
of the previous posts that the nodes run NFS.  I tried SSHing to several
nodes and checking the time on each node and noticed the nodes vary by a
few seconds from each other and they differ ~4 seconds from the console.
Is there a way to force an NFS time sync on the nodes?  Or does the
Orbit library provide some time sync facility to be able to correlate
events?

Thanks,