perfSONAR User Q&A and Other Discussion

[Andrew Lake <>]

Hi Bill,

Looking back at your issue, the original error reported in November got fixed, but looks like you reported a different error in April under that same issue after it was closed. Probably how it got lost in shuffle. I created a new issue just now for that latest error: https://github.com/perfsonar/mesh-config/issues/95. I’ll throw it in our sprint that starts next week and have a look. Looks like it is probably mesh-config related as opposed to pscheduler related but we’ll dig in and find out. Sorry for the confusion.

Thanks,

Andy

On June 30, 2017 at 10:44:58 AM, Bill Owens () wrote:

Andy,

 

I apologize if I misunderstood the notes; from my reading it appears that Mark Feit was unable to reproduce the bug and closed it as though it did not exist. I believe that I was able to reproduce it much later and appended that to the issue, as well as emailing the list. I haven’t seen any change in the behavior of my pS machines since then.

 

Bill.

 

From: Andrew Lake <>
Date: Friday, June 30, 2017 at 10:25 AM
To: "" <>, Bill Owens <>, Charley Kneifel <>
Subject: Re: [perfsonar-user] perfSONAR v4 and IPv6

 

Hi Bill,

 

No one’s ignored that issue, that issue has not been observed as far as I know since that early RC when you reported it. If someone actually see’s it we are happy to re-open it. That was a simple JSON validation issue that was resolved during one our our early RCs and does not encapsulate every possible thing that could go wrong when someone tries to use IPv6. Any IPv6 issues observed since then have been completely different problems, and we have committed a number of IPv6 related fixes since then. If there are any additional issues people are seeing using IPv6 we want to know about it, but I suspect they will be new issues and not have anything to do with #233 that you shared. 

 

Thanks,

Andy

 

 

 

On June 30, 2017 at 8:32:23 AM, Bill Owens () wrote:

Have a look at https://github.com/perfsonar/pscheduler/issues/233 and see if that matches your experience. If it does, perhaps you’ll have more luck at getting someone to pay attention to it…

 

Bill.

 

From: <> on behalf of Charley Kneifel <>
Date: Thursday, June 29, 2017 at 9:13 PM
To: "" <>
Subject: [perfsonar-user] perfSONAR v4 and IPv6

 

Hello All,

 

I am unable to get IPv6 tests to work.

 

I can ping a remote IPv6 node:

 

ping6 sdsc-pt1-v6.es.net
PING sdsc-pt1-v6.es.net(sdsc-pt1-v6.es.net) 56 data bytes
64 bytes from sdsc-pt1-v6.es.net: icmp_seq=1 ttl=48 time=107 ms
64 bytes from sdsc-pt1-v6.es.net: icmp_seq=2 ttl=48 time=106 ms
64 bytes from sdsc-pt1-v6.es.net: icmp_seq=3 ttl=48 time=105 ms
64 bytes from sdsc-pt1-v6.es.net: icmp_seq=4 ttl=48 time=104 ms
^C
--- sdsc-pt1-v6.es.net ping statistics ---
4 packets transmitted, 4 received, 0% packet loss, time 3726ms
rtt min/avg/max/mdev = 104.495/105.886/107.681/1.283 ms

name services are resolving properly

[root@ps-node-c pscheduler]# nslookup -type=aaaa sdsc-pt1-v6.es.net
Server:         192.168.202.5
Address:        192.168.202.5#53

Non-authoritative answer:
sdsc-pt1-v6.es.net      has AAAA address 2001:400:1101:1150::3

 

I have added an IPv6 host to a PS4 node and confirmed that it is set for IPv6

 

I see no tests with the hostname (sdsc-pt1-v6) in the pscheduler logs and no tests with --ip-version=6 in the logs either.

 

I do have a private v4 address, but have a public V6 address that is exposed for icmpv6.

 

Anyone have any suggestions?

 

Charley