June 3, 2019
CF2018 PMT connectivity issues
Comments
(13)
June 3, 2019
CF2018 PMT connectivity issues
Newbie 1 posts
Followers: 0 people
(13)

we’re in a segregated enterprise network environment.  i’m trying to figure out how to get my workstation connected to our dev server.  I’m running PMT on localhost, I go to Settings and CF servers and add a new manual entry.   I put in the DNS name for the dev server and port 8500.  but when I submit changes, it attempts but fails to connect to the remote node.    I can open http://dev-server:8500/CFIDE/Administrator/index.cfm  from a browser on my workstation and get the CFIDE login.

the error PMT reports is “Access Forbidden”

i also tried to add one of the extra instances we have on the dev server.  I get a different error “Not able to communicate with ElasticSearch instance host: DEV-YR and port:9,250, hence Monitoring Service is not initialized.   but that’s …. my local workstation.  and datastore service is running.  in fact, PMT is monitoring my local CF instance just fine.

13 Comments
2020-09-29 02:41:19
2020-09-29 02:41:19

As a follow-up comment on all the issues discussed here, I’ll point out that there was a helpful webinar from Adobe engineer Nimit Sharma who reviewed in much more detail the configuration of the PMT, and the connection/communications between the CF instance the Datastore, and the PMT service itself:

https://cfwebinar-pmt-architecture.meetus.adobeevents.com/

Anyone can register to get access to the on-demand video of about an hour. Lots of great details offered.

Like
2019-06-07 09:23:23
2019-06-07 09:23:23

Hi,

For the first issue which you are mentioning, please copy the secret from Dev server admin’s PMT page and copy it while registering it with PMT as suggested by Charlie.

For the second issue it seems like your second instance is not able to communicate to ElasticSearch of PMT  at host: DEV-YR and port:9,250.

For further support please reach out to cfinstal@adobe.com or please share your mail id we will setup a screen share session.

Thanks,

Mayur Jain

Like
(3)
>
MayurCF
's comment
2019-06-07 11:00:06
2019-06-07 11:00:06
>
MayurCF
's comment

Mayur, thanks for jumping in. As for your second point, is it really so that the CF instance being MONITORED needs to reach the datastore? I’ve assumed (as many may) that it would be the PMT only which talks to the datastore, having pulled metrics from the CF instance. Can you clarify?

Like
>
MayurCF
's comment
2019-06-10 05:58:49
2019-06-10 05:58:49
>
MayurCF
's comment

Hi Charlie,

Actually Every CFinstance pushes the date to the datastore. On registering a CF node with PMT. PMT shares the datastore details with CF node, which is then used by CF to push metric data to datastore. PMT can talk to individual CF nodes for active request monitoring, thread and heapdump.

Thanks,

Mayur Jain

Like
>
MayurCF
's comment
2019-06-10 12:43:25
2019-06-10 12:43:25
>
MayurCF
's comment

Got it. Thanks.

Is there any page (in docs or a blog post) that depicts that route of communication among the 3 components (instance, pmt, datastore), to help folks anticipate/confirm successful communication when things are not workimg? Sorry if I’ve missed or forgotten it.

Like
2019-06-05 14:57:51
2019-06-05 14:57:51

I’m having exact same issue; main CF instance + PMT works fine, but it won’t let me manually add a second server without the same error message (except mine says localhost, 9250 – and there IS something running locally on 9250, this ElasticSearch thingy I imagine). But there isn’t anything running on port 9250 of the “other” server to be monitored, so opening that port won’t have any effect.  The error message appears to be extremely misleading.  The interface for adding/changing the servers to be monitored doesn’t look very promising either, e.g. there’s not even a way to delete a server once you’ve tried adding it – you can only enable/disable it !

Like
(1)
>
AndyLorenz
's comment
2019-06-06 13:39:03
2019-06-06 13:39:03
>
AndyLorenz
's comment

Andy, there is indeed a lot about the PMT that can be hard to setup, as there are so many moving parts–and all the more when the PMT and (its datastore) are on separate machines from the server(s) to be monitored.

Yes, of course, that’s how it’s DESIGNED to work. I’m just saying that people do struggle when some aspect of configuration doesn’t work. There are multiple diagnostics to consider, as well as multiple configuration options to possibly tweak. Too many to easily propose what to do next.

And as I reported to the op (yuliangr), that complication may call for reaching out to Adobe directly to get their help, at cfinstal@adobe.com. There’s just no guarantee they will see these discussions here.

Like
2019-06-04 15:55:57
2019-06-04 15:55:57

Are you putting the secret key (shown in the Dev svr cf admin pmt settings page) into the pmt pointing to it?

Like
(5)
>
Charlie Arehart
's comment
2019-06-04 18:21:11
2019-06-04 18:21:11
>
Charlie Arehart
's comment

yup definitely have the secret key entered.

I think possibly what I’m missing is allow traffic on port 9250 from monitored node to PMT ES instance.  waiting for IT to verify that’s what’s needed.    the setup guides seems to assume that PMT and monitored node can always communicate

Like
>
Charlie Arehart
's comment
2019-06-04 18:34:48
2019-06-04 18:34:48
>
Charlie Arehart
's comment

interesting too….apparently the secret key is attached to the license somehow.  i just applied our enterprise license to the dev server and the secret key changed.

Like
>
yuliangr39032863
's comment
2019-06-04 18:36:36
2019-06-04 18:36:36
>
yuliangr39032863
's comment

Interesting stuff. Do keep us posted. 🙂

Like
>
yuliangr39032863
's comment
2019-06-04 18:40:40
2019-06-04 18:40:40
>
yuliangr39032863
's comment

I had IT open the firewall from dev server to my workstation port 9250.   retried the PMT connection.   it thinks for a bit longer and then returns hard error.    I’m installing wireshark to see who is actually throwing this error, the pmt or the monitored node.

HTTP Status 500 – Internal Server Error


Type Exception Report

Description The server encountered an unexpected condition that prevented it from fulfilling the request.

Exception
java.lang.NullPointerException
com.fasterxml.jackson.core.JsonFactory.createParser(JsonFactory.java:879)
com.fasterxml.jackson.databind.ObjectMapper.readValue(ObjectMapper.java:2842)
coldfusion.monitor.es.JSONConverter.toMonitorSettings(JSONConverter.java:164)
coldfusion.monitor.es.Configuration.getMonitorSettings(Configuration.java:482)
coldfusion.monitor.es.Configuration.initializeSettings(Configuration.java:388)
coldfusion.monitor.MonitoringServiceImpl.initialize(MonitoringServiceImpl.java:662)
coldfusion.monitor.MonitoringServiceImpl.store(MonitoringServiceImpl.java:722)
coldfusion.monitor.MonitoringServiceImpl.intializeNodeWithEs(MonitoringServiceImpl.java:501)
coldfusion.monitor.PMSGenericServlet.doPost(PMSGenericServlet.java:194)
javax.servlet.http.HttpServlet.service(HttpServlet.java:660)
javax.servlet.http.HttpServlet.service(HttpServlet.java:741)
coldfusion.bootstrap.BootstrapServlet.service(BootstrapServlet.java:311)
org.apache.tomcat.websocket.server.WsFilter.doFilter(WsFilter.java:53)

Note The full stack trace of the root cause is available in the server logs.


Apache Tomcat/9.0.10

Like
>
yuliangr39032863
's comment
2019-06-04 18:51:49
2019-06-04 18:51:49
>
yuliangr39032863
's comment

That’s an odd one. Use of the pmt seems low, judging from community feedback here, so I’m not aware of that error or it’s cause, or even best steps to diagnose.

If you don’t solve it, email the folks at cfinstal@adobe.com, as it’s about setup. And I hope you’ll report back with the resolution, so all can learn.

Like
Add Comment