perfSONAR User Q&A and Other Discussion

[Winnie Lacesso <>]

Greetings Maestro!

On Mon, 2 Feb 2015, Aaron Brown wrote:

> Hey Winnie,
> 
> A reboot is probably easiest, but you should only need to restart
> 'regular_testing'. The rest should be fine. 

Agreed about reboot - will do later; but last time "just a reboot" ended 
up in PAIN with 0% free inodes & a disk full of a zillion tiny files.

Anyway decided to start them in batches & watch for failures. After 
successful start of perfsonarbuoy_ma & perfsonarbuoy_owp_collector, the next 
one
failed:

root@lcgnetmon>
 service perfsonarbuoy_owp_master start
Conf::must_get_val(ATTR=>NODE, ) undefined, called from 
/opt/perfsonar_ps/perfsonarbuoy_ma/bin/powmaster.pl:154
Couldn't start perfSONAR-BUOY OWAMP Measurement Service

Can someone help translate where to find its problem?

Something is writing to owamp_bwctl.log

Feb  2 15:52:40 lcgnetmon owampd[14270]: FILE=sapi.c, LINE=466, 
ControlSession([lcgnetmon.phy.bris.ac.uk]:861) accepted from 
userid(nil):([epgperf.ph.bham.ac.uk]:56242)
Feb  2 15:52:40 lcgnetmon owampd[11229]: FILE=policy.c, LINE=1811, ResReq 
ALLOWED: regular:request:bandwidth = 3360 (result = 208320, limit = 
10000000)

& something has spawned 148 of these:

owamp    13926 13925  0 15:50 ?        00:00:00 /usr/bin/owampd -c 
/etc/owampd -R /var/run

That's on the BROKEN Latency box; the other Bandwidth box has 1 (not 148) 
of those owampd processes. Are many of those on a Latency box normal?


(Haven't re-enabled the crontabs yet. Last.)