perfSONAR User Q&A and Other Discussion

[Andrew Lake <>]

Hi,

My guess is you are hitting the same issue as the users in a thread earlier in the week that could not save their test results. They were able to fix this by doing the following:

1. Open /etc/perfsonar/toolkit/configdaemon.conf and change “localhost” to “127.0.0.1”.

2. Save the file and restart the configdaemon with “systemctl restart perfsonar-configdaemon” or “/etc/init.d/perfsonar-configdaemon” depending on whether you have CentOS 7 or CentOS 6 respectively.

The problem appears to be that the configdaemon only listens on ::1, but the web backend is trying to connect on the IPv4 address. Seems to be an inconsistency on  how the localhost lookup is performed. We will have a patch in our next update, sorry for the inconvenience. 

Thanks,

Andy

On May 18, 2018 at 1:46:32 PM, Rolf Seuster () wrote:

Hi all,

I don't understand this - I am not able to update the "Administrative
Information" of
my perfsonar box.

For tests I have two boxes, both with IPv4 and IPv6.
ps-bandwidth.hepnetcanada.ca
participates in ATLAS and Belle II meshe, the other box runs only local
tests.

On both machines, I am not able to update the administrative
information. After
I put in all fields and press "save", all info just typed in disappears
after a short moment.

The "interesting" thing is, that in
/var/log/perfsonar/web_admin/web_admin.log
the last lines seem to indicate that this host is found in some database:

2018/05/18 10:43:47 (192267) DEBUG> Utils.pm:119
perfSONAR_PS::Client::Utils::send_http_request - Sending HTTP GET to
http://ps-west.es.net:8090/lookup/records/?type=host&client-uuid=38921940-4756-11E7-AAD5-E8E1B000BEC5
2018/05/18 10:43:47 (192267) ERROR> Host.pm:273
perfSONAR_PS::NPToolkit::DataService::Host::get_details - Found host
record in LS using 206.12.154.60

wgetting this entry indeed shows that e.g. the host-administrators are
not set.

Confused, because according to the perfsonar docs, the update should be
trivial.

Thanks for any pointers, Rolf