perfSONAR User Q&A and Other Discussion

[Shawn McKee <>]

Hi Winnie,

It shouldn't hurt to do a reboot.   For the latency node to begin testing it needs to have its clock synchronized.  Depending upon how your NTP is configured this can take a little while.

We are monitoring your instances here: https://maddash.aglt2.org/WLCGperfSONAR/check_mk/index.py?start_url=%2FWLCGperfSONAR%2Fcheck_mk%2Fview.py%3Fview_name%3Dhosts%26host%3Dbri

(You need an x509 certificate from an IGTF CA to be able to view this)

Right now the instances don't seem to be reachable.  Maybe a reboot may help.

Shawn

On Mon, Aug 3, 2015 at 9:55 AM, Winnie Lacesso <> wrote:

Bonjour,

Thank you VERY much for your advice about this. A great relief.

I wrote;
> A tiny concern, /var/log/perfsonar/regular_testing.log is logging what look
> like errors:
>
> 2015/07/31 11:51:52 (31039) ERROR> MeasurementArchiveChild.pm:125
> perfSONAR_PS::RegularTesting::Master::MeasurementArchiveChild::__ANON__ -
> Problem handling test results: Problem storing results: Error writing
> metadata: 500 Can't connect to localhost:80 (connect: Connection refused)
> at

> As it happens the room they're in is losing power tonight so they'll be
> shutdown by 16:00 today & rebooted Monday morning.

They rebooted this morning but there was initially an issue with network
equipment so they had no network. Finally when ethtool eth0 showed they
had a live link, service network restart -> up ok with 1Gbps link; but I
wonder if they should be rebooted once more so everything starts as it
should.

On Fri, 31 Jul 2015, Andrew Lake wrote:
> Let me know if the error in regular_testing.log persists on reboot. It

They do not. However, in looking at graphs btw 137.222.74.15 & other
hosts, it's showing 100% packet loss (solid pink line), & graphs btw
137.222.74.17 & other hosts show solid red 1 "Error"

Might it just take some time for data from them on new subnet to settle
down?
Am still thinking one more reboot without network badness might be
best (all starts healthily)....