We seemed to be having issues within our environment but I don’t know where to start troubleshooting.
I created a fixlet that has no relevance and is subscribed to all sites and all endpoints. So it should becoming relevant to 10k machines in a matter of minutes. I created a new test fixlet (as well as others) and within 4 days none of these have any applicable computers. Where should I start to look for this kind of an issue?
Health checks indicate all is well and our relays are supposedly working as well.
Relevance is just “all computers” and to ensure the test fixlet had full access I put it in master action. Every fixlet\task\baseline we’ve created in the last week is acting the same, regardless of site or operator.
i do agree with @paulochoi.
If you want to deploy the fixlet on all the endpoints select “All computer” and for writing relevance expression, it should must be TRUE instead of “no relevance”.
Hope it’ll works.
We’ve done "all computers’ before with many items as it is a radio button amongst the other options. It has always worked for us in the past. Just to be sure, I did select the default ‘True’ relevance and I’m seeing the same results. I’m creating the fixlet as a master operator in the master action site. Confused as to what happened in the last week. All of this was reporting in just fine a week ago and we haven’t changed anything.
Even though there are no applicable computers, we can still run this task against any endpoint just fine.
I’d say there is something wrong with your Master Action Site propagating, but that doesn’t seem to be the case if taking action works.
Also, if taking action works, then that should mean that there is nothing wrong with your licensing. Does taking action work from custom sites as well as the Master Action Site?
Are reports flowing from clients to relays to root server?
Also, does the fixlet/task in question show up on any clients? Should be an FXF file with the ID of the item in question on the client.
Yes the actions are reporting their action status from custom and master site. I attempted to push a custom test action from master action and used baretail to monitor the logs. I never saw this action hit my PC and to answer your initial question, no, the action is not located in the fixlet doesn’t show up on my client as an FXF file. I did a refresh of my endpoint and it instantly refreshed all the opsites, but none of the custom sites (i don’t know if a refresh populates all sites or just opsites?).
So that got me thinking and I recreated the custom test action with relevance of ‘True’ within my own opsite only. Applicable computer count still does not show up, but this time the local logs do show the information coming through. It reports as “not relevant” but at least shows that it received the command. Checking the local opsite folder on my computer, the FXF file is not present.
This is an odd one. Definitely something for IBM Support, as I’m not sure how to help from here.
We did have something odd happen to us where custom sites that had not had any changes made to them recently would not propagate to the relays after a relay upgrade or something like that. We had to make a change to every custom site to get it to propagate, which fixed an issue a bit like this. This would not be relevant if the sites you are working with have been modified in some way, and it did not affect the master action site.
I would check your relays starting from the top and working your way down to see how things are propagating. I would also wonder if you create a task with relevance of TRUE and no actionscript and take it on the root server if that works, then do it to relays and see if that works, and work your way down.
No, I have the console installed on all relays and I’ve tested different content as different logons, sites, subscriptions, from various consoles. But yes I did clear the cache multiple times on each console.
Yea I ended up contacting support and getting in touch with Jgo. We’re at a sev 1 now and hopefully someone will be contacting me shortly to look into this. He thinks it may be a database issue but we really have no idea yet. Appreciate everyone’s responses on this!
Followed these steps and everything started working. odd thing was, we upgraded from 8.x several months ago, so we don’t know why all of a sudden this became an issue.
