perfSONAR User Q&A and Other Discussion

[Joshua Baker-LePain <>]

We're having some issues getting our perfSonar node to pull down the mesh 
configuration, and I'm about to start pulling out what's left of my hair. 
Our pS node is running up-to-date CentOS-6 and the most recent perfSonar 
RPMs (4.0-1).  The node is dual-homed, with eth1 on our private network. 
The routing looks like this:

Destination     Gateway         Genmask         Flags Metric Ref    Use Iface
172.26.1.0      172.26.2.1      255.255.255.192 UG    0      0        0 eth1
172.26.2.0      0.0.0.0         255.255.255.192 U     0      0        0 eth1
169.230.239.0   0.0.0.0         255.255.255.0   U     0      0        0 eth0
172.19.0.0      172.26.2.1      255.255.0.0     UG    0      0        0 eth1
0.0.0.0         169.230.239.1   0.0.0.0         UG    0      0        0 eth0

In /etc/perfsonar/meshconfig-agent.conf, the mesh is defined as:

<mesh>
        configuration_url https://ps-dashboard.cenic.net/CENIC-Dashboard.json
        validate_certificate 1
        ca_certificate_file          /etc/pki/tls/certs/ca-bundle.crt
</mesh>

But everytime I run 'service perfsonar-meshconfig-agent start' this is 
what I get in meshconfig-agent.log:

2017/05/12 11:58:17 (21950) ERROR> Agent.pm:302 
perfSONAR_PS::MeshConfig::Agent::__configure_host - Problem with mesh 
configuration: 500 Can't connect to ps-dashboard.cenic.net:443 (): 500 Can't 
connect to ps-dashboard.cenic.net:443 ()

I can manually pull down 
<https://ps-dashboard.cenic.net/CENIC-Dashboard.json> via wget or a perl 
one-liner.  I've even tried shutting down eth1 just to make sure there's 
no weird routing going on, and it made no difference.  I verified via 
wireshark that there is absolutely no traffic between our host and 
ps-dashboard.cenic.net when I start the meshconfig-agent -- it's not even 
trying to make contact.

Any help would be greatly appreciated.  And please let me know if I left 
out any relevant info.  Thanks.

--
Joshua Baker-LePain
QB3 Shared Cluster Sysadmin
UCSF