[SCore-users-jp] [SCore-users] 5.8.x

[Nick Birkett]

I have noticed this error happens quite often
when using scout -F hosts -e scrun -nodes=n ./myexec.exe

<0> ULT:ERROR Hostinfo recv error (length)

The <0> is not significant - it can be any number.

When this happens the job does not start. However
if we kill the scout process (using cntrl C or pkill)
and run the job again it works (using same hosts file).

I have seen this on RedHat 7.3,9,Opteron Suse 9 using
Score 5.8.1 (the one originally developed for Suse
9 Opteron) and also 5.8.2.

If the job is submitted with a job scheduler then the user
thinks the job is running when it isn't.

I can't remember seeing this on Score 5.6.1.

Is this a know bug ?

Nick
_______________________________________________
SCore-users mailing list
SCore-users ＠ pccluster.org
http://www.pccluster.org/mailman/listinfo/score-users