perfSONAR User Q&A and Other Discussion

[Andrew Lake <>]

Hi,

The old centralmanagement was the combo of esmond and maddash. We now have the “perfsonar-archive” package which handles getting OpenSearch installed and configured for you (i.e. this is the new equivalent esmond). See https://docs.perfsonar.net/multi_ma_install.html for info on running a new central archive. You can just run “yum install perfsonar-archive maddash”  and get the same result the old perfsonar-centralmanagement. 

Hope that helps,

Andy

On April 18, 2023 at 2:40:30 PM, Jarett DeAngelis () wrote:

Hi Andrew,

Just for clarification — what replaces the role the perfsonar-centralmanagement package formerly occupied in pS5?

Thanks,

Jarett

On Apr 18, 2023, at 12:47 PM, Andrew Lake <> wrote:

Apologies, my mail client seemed to mess with the URLs in the various commands I shared (removed https://). Trying again….the correct commands below:

- To check if you are affected by this issue run “yum info perfsonar-centralmanagement”. If the version starts with “5.0.0” then you got hit with this. If its still 4.X then you can go ahead and close this email. 

- To reverse this change, run the following commands to remove the new packages and reinstall Esmond with all your old data in tact:

yum downgrade https://software.internet2.edu/rpms/el7/x86_64/4/RPMS/perfsonar-centralmanagement-4.4.6-1.el7.noarch.rpm

yum remove perfsonar-archive perfsonar-logstash perfsonar-logstash-output-plugin perfsonar-elmond logstash-oss opensearch

yum reinstall https://software.internet2.edu/rpms/el7/x86_64/4/RPMS/esmond-4.4.6-1.el7.x86_64.rpm

systemctl restart httpd cassandra

- Finally, to check that esmond is working again, run the following command and if you get a bunch of JSON, then Esmond is good to go. If not, review the steps above or feel free to reach out to this list:

curl -k "https://localhost/esmond/perfsonar/archive/?limit=10"

On April 18, 2023 at 1:24:25 PM, Andrew Lake () wrote:

All,

For those running the perfsonar-centralmangement hosts, the intent with 5.0 was to not auto-upgrade your central Esmond instances. Unfortunately, it looks like an old beta central management 5.0.0 beta package got copied into our main RPM repo that did not honor this. The package has been removed from our repo but some of you still may have gotten hit by this. Good news is that it is completely reversible with a few commands. Follow the steps below:

- To check if you are affected by this issue run “yum info perfsonar-centralmanagement”. If the version starts with “5.0.0” then you got hit with this. If its still 4.X then you can go ahead and close this email. 

- To reverse this change, run the following commands to remove the new packages and reinstall Esmond with all your old data in tact:

yum downgrade https://software.internet2.edu/rpms/el7/x86_64/4/RPMS/perfsonar-centralmanagement-4.4.6-1.el7.noarch.rpm

yum remove perfsonar-archive perfsonar-logstash perfsonar-logstash-output-plugin perfsonar-elmond logstash-oss opensearch

yum reinstall https://software.internet2.edu/rpms/el7/x86_64/4/RPMS/esmond-4.4.6-1.el7.x86_64.rpm

systemctl restart httpd cassandra

- Finally, to check that esmond is working again, run the following command and if you get a bunch of JSON, then Esmond is good to go. If not, review the steps above or feel free to reach out to this list:

curl -k "https://localhost/esmond/perfsonar/archive/?limit=10”

We are really sorry for the inconvenience this may have caused and please let us know if you have questions.

Thanks,

Andy

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