perfSONAR User Q&A and Other Discussion

[Tim Chown <>]

Hi,

Raul’s observation of the behaviour is captured in https://github.com/perfsonar/psconfig/issues/112.

 

Tim

On 27 Sep 2023, at 16:46, Phil Reese <> wrote:

Thanks Christopher and Raul!

Here is an additional twist and it may be some help pinpointing the issue.

Both of you seem to be talking about issues with a full Toolkit host, and your debug approaches make sense.

In my case, my issue is with with the Archiver v5 talking to a MaDDash v5 host.  The implication is that while I seem to have the same issue with archiving stopping, I have many fewer services!

My failing host only has these three pS5 rpms loaded, perfsonar-archive, perfsonar-psconfig-publisher and maddash.

From Raul's list I'm only running these services, httpd, opensearch, logstash, and elmond.

I guess it might be possible that my testpoints aren't forwarding results to the archiver, but I'm archiving to an older esmond process and to the newer opensearch.  The elmond MaDDash based server continues to show appropriate data. 

Phil

PS- Since I've been testing pS5 for some time I've an Ansibile yml file which rebuilds my servers very quickly, which I've done and data is appearing again.  I'll be tracking how long the archiving continues just for another data point.  Looking forward to v5.0.5!

On 9/27/23 5:44 AM, Raul Lopes wrote:

Hi Phil,

 

I worked on a couple of hosts yesterday, where increasing the shards didn’t solve the problem. I had to stop all perfsonar services (see [0]) and start them back on.

 

Regards,

Raul

[0]

 

 

systemctl stop httpd

systemctl stop owamp-server

systemctl stop twamp-server

--
To unsubscribe from this list: https://lists.internet2.edu/sympa/signoff/perfsonar-user