Is anyone updating their BESAgents for RHEL/CentOS outside of BigFix? We are hoping to use both BigFix and a yum repo to offer updates but have hit a road block.
We are looking to utilize a shared repo to make BigFix available as a fresh install and update for hard-core early adopter Linux lovers. BUT prior to removing the old agent, the BESAgent service is stopped, and is not started again after the old agent has been removed. This leaves the system in a state of having the latest version of BESAgent, but does not start the agent automatically.
We could modify the RPM postun section to start/restart the service, but that adds additional overhead for each upgrade. In addition future upgrades pushed via BigFix would negate the customizations, unless we overwrote the default installer.
Is there any way to use BOTH BigFix & a RHEL Repository to update endpoints, or are we stuck with one or the other?
It’s not any help to you RIGHT NOW, but there was an RFE recently (Request For Enhancement) to cover a “Client Helper” component similar to what is available under Windows but for Linux systems. One of the features of that helper would potentially be to restart a stopped BES Client process.
If you have not done so already, search the Forums to find the thread about it and use the link in there to go “Vote” for the RFE. The more Votes and RFE gets the greater the chance that IBM will devote resources to making it reality.
So I’m confused why you would want to update the agent outside of BigFix itself as the fixlets do a better job? Some operators definitely don’t like having “new” agents running in a system (as there are other fixlets that can warn if the agent is toooo new) but it generally does work. And as these are mostly one off systems then just downloading and running the RPM would be easy no?
Just my opinion here, but the OS does a much better job of restarting failed services than we can. I’d actually like the existing Windows BESClientHelper to go away (as its so old) and the new OS’s can restart a service if it fails.
I get the issue if someone stops the service manually but I think there are ways we can stop that from being done as well (I’ve seen services I could not stop) which we could make available by content.
The main reason why we are looking at this is to make die hard Linux admins happy. If they can pull it from a repo before we push it to them via BigFix it works for everyone.