NetBSD's reliability [Was: raidframe in netbsd-5?]
Matthew Mondor <[email protected]>
| Newsgroups | gmane.os.netbsd.advocacy,gmane.os.netbsd.current |
|---|---|
| Message-ID | <[email protected]> |
On Thu, 13 Jan 2011 12:06:12 -0500 Greg Troxel <[email protected]> wrote: > And a hat tip to releng@; it's really impressive that this is the first > time in probably 6 years when I was not comfortable to just "update > along stable branch and reboot, no worries", and even this is very minor > and apparently an odd case. I'll go read the diff when I have a chance. I concur, the stability and reliability of NetBSD (other than its versatility) is a major reason I've kept with it, and there are some systems which are still running which I upgraded from source regularily over the years, two of them old enough to have started with NetBSD 1.6, most of them now tracking the netbsd-5 branch and now running NetBSD 5.1_STABLE. Most of those systems have been upgraded remotely via SSH using NFS or a copy of the re-built source tree (built using a dedicated host), after testing the procedure on a local box running the same NetBSD version. I've also often moved the software/data around to new hardware instead of reinstalling, for systems I have physical access to. The binary compatibility for older releases also served me well until the packaged and custom software could also be upgraded or rebuilt, and for booting+testing kernels before userland upgrades. Since I started using NetBSD (in the 1.5 days, comming from Linux and exploring alternatives) I could say that I've seen very few upgrade glinches; off the top of my head I can now remember three, one related to a cgd configuration bug, one a postinstall script bug missing a pf related group, and a PAM and /etc/login.conf related problem, which have been promptly fixed. I can remember a few X11 related issues here and there on local workstations, as well as some power management ones, notably after switching from APM to ACPI and XFree86 to XOrg, but no other open source Operating System I know was spared in those areas either, and those problems didn't affect the headless systems. Indeed, thanks for consistently keeping NetBSD a high-quality OS for all those years. (I hope you don't mind me CCing netbsd-advocacy :) -- Matt