Beware the spacewalk
I think spacewalk is great. It’s a great product provided for free and sponsored by Red Hat. In fact, I think it’s so great that it’s going to keep RHEL-based distros relevant for non-enterprise users for the next 2-3 years. Of course, then someone will add support for debian/unbuntu, but that’s beside the point.
I think spacewalk is great. I’ve been using it to good success for a little while now. It hasn’t reached 1.0 yet, but that hasn’t given me a single problem. Until now.
I expected the roadmap to 1.0 was more of a process of removing the Red Hat proprietary bits, removing Oracle dependence, finishing the migration to java, adding the open source flourishes, etc. The releases up to this point have been reliable and robust.
Then Spacewalk 0.7 released a version of rhnsd that was broken. Not just a little bit broken, either. A lot broken. Broken in a way that means it doesn’t work under any circumstances. Broken in a way that means the code was never tested prior to commit or release. Also it leaks file descriptors. Also it leaks memory.
rhnsd isn’t a minor part of the client. It’s the daemon that runs in the background and calls rhn_check on a configurable interval. If rhnsd isn’t working, that means clients don’t check in. If clients don’t check in then there’s really no point in having Spacewalk.
I expect there to be bugs in pre-1.0 software, of course, but I don’t think it’s too much to ask that code at least be trivially tested prior to release. I still like Spacewalk, but it’s clear that I should be a lot more active in my support of the project.
The thing I don’t understand, though, is that Spacewalk has been released for a month. How has no one else noticed a problem? The community seems pretty active. Am I using Spacewalk in a novel way? Does no one else upgrade their clients? It just doesn’t add up. Maybe I’m doing it wrong.