I am new to creating custom plugins and actually attempted to use hyperic support before posting here. If anyone has access to the hyperic Jira at http://jira.hyperic.com, my ticket reference is SUPPORT-3321. We worked on the issue for over 2 weeks before the lady helping me suddenly realized that I didn't have custom plugin support. Regardless, the support given was of very little use. My situation now is that I have a custom plugin that loads through the UI when it is in /home/hyperic/hq-plugins, but does not display metrics. The lady helping me told me it should be in /home/hyperic/agent-3.1.4/pdk/plugins/ only, but that caused stack trace errors (PluginNotFoundException) in the UI. I was wondering if some one could take a look at the attached custom plugin and the associated MIB file and possibly tell me what I may be doing wrong? When I run:
[DEBUG] MIBTree - Loading MIB: '/usr/share/snmp/mibs/product_nozero.mib.txt': success [DEBUG] PluginDiscoverer - Set platform.type=sotechProductLS [DEBUG] PluginDiscoverer - No detector for product [DEBUG] PluginDiscoverer - Adding Auto Scanners= [DEBUG] Scanner - Running discovery for another platform: sotechProductLS=localhost/127.0.0.1 [WARN] Scanner - No server signatures were found. [DEBUG] Scanner - Running discovery for another platform: sotechProductLS=localhost/127.0.0.1
It would be nice to see server and agent logs what this plugin is actually trying to do.
3.1.4 is rather old and at least I havent never used it. In version 4.x this plugin wont deploy until classpath setting in xml is moved from platform section to plugin section. Don't know why, perhaps 3.1.4 has different xml validation. Also in version 4.x, default enabled mertics for this plugin has collection interval of 30 minutes, so in any case it would take some time to collect any metrics. But I guess you have changed those since you were able to deploy the plugin.
I belive JIRA doesn't show others support tickets, so it's difficult to know what you have discussed with hyperic.
I'm about to do my own custom plugins for snmp, so this case sounds interesting.
Does it really worth it that much to purchase Virtual Center? Would it be worth it to purchase Virtual center If I don't upgrade my VI3 from standard to Enterprise. The post above cheered me up a lot into buying it, but it did not mention what it will give & not give in standard edition.
I noticed that error and began to research what it should be for 3.1. Unfortunately there is no information for 3.1 in the SNMP plugins section. I looked at 3.2 and determined what you have here to be correct. So I replaced my entries as listed below:
Could you post your lates sotechProductLS-plugin.xml. I may want to try this myself.
Btw, those errors in server log looks pretty nasty. If you remove this plugin from server, does those errors still occur? Are you in position to wipe you HQ installation, because that maybe something what you need to do. (hope this is your testing instance)
I looks like there is something in the database what should not exist.
I attached my latest version of the plugin. This is my lab instance and I am prepared to blow it away if need be after testing. However, the server log looks good for the most part when I rename the plugin to .new, though a temp file throws an error on that file in the deploy directory.
1. Those unzip errors are not related to this. You propably have some extra files what it's trying to deploy.
2. This loads to my 4.x instance without errors. So, I would suggest you to try this on clean HQ installation.
3. I believe there's something what needs to be added to xml file:
I would use <config include="snmp" />. I don't exactly know what that include, but you may have already manually added those configs.
I think server section also needs this <plugin type="measurement" class="net.hyperic.hq.product.SNMPMeasurementPlugin"/> definition. If you later want get metrics from there, I think.
Last what comes to my mind, is filters. I think snmp measurement wont work until some filters are introduced(you see these within every plugin). I've seen those used in snmp plugins. At least you are setting template on your metrics, so this may not be that big issue.
I had a plan to do plugin for our printers and ip-phones. I don't need it, but my colleague wants to have it.
But, first you need to be able to load your plugin without errors.
That has me a lot closer than I have been so far. I still get some nasty DB errors in the server log, so I am probably going to have to reinstall to resolve that. Here are the latest plugin and log files. I will probably wipe the db today and let you know how it loads on a new install.
2008-12-23 10:19:26,459 INFO [main] [org.hyperic.hq.product.server.session.ProductManagerEJBImpl] sotechProductLS unknown -- registering
When I tried to deploy your plugin, I also got that line. I forget to copy those mibs to that /usr/share/snmp/mibs directory on server side. Once I did that, plugin deployed without errors. You propably have mibs in place, just wanted to tell you this.
Yes, sotech... is now available under Monitoring defaults | Platform types. I am also able to add it through browse resources | tools | new platform | sotech... However, I have been able to do that for some time. The issue is that I never get any metrics. The icon or availability indicator is always a grey ? and never a green check. I have waited for over 30 minutes on every occasion that I have created it and never see the status change.
You know, I don't know how that older hq version is working but there should be some kind of error somewhere. For the UI is showing error because from the agent I don't have any snmp devices to connect to.
Are you really seeing empty metric table. I mean empty table where those merics are just unavailable? Or do you mean that completely page is blank?
I think this could work If my agent would find matching snmp device with correct values.
What comes for the availability, your xml is missing metric for availability. As you can see it from 'monitoring default' page.
I think you are getting pretty close to get this working 🙂
When I select the platform that the plugin is assigned to, no metrics display. It states "There are no indicator metrics to display." I added the filter and the availability metric. My latest plugin is attached along with logs. I am going to blow away my DB to see if that will resolve the remainder of the errors in the server.log.
Okay, after deleting my DB and starting over everything loads without error using the attached plugin. However, when I try to add the plugin now through browse resources | tools | add platform, I receive the following error:
The configuration has not been set for this resource due to : Invalid configuration: Invalid request: Failed to lookup OID for name=usageChange (last lookup failure=null)
I thought my metrics maybe programmed incorrectly. I am going to look that one up and try a few things.