On 2/17/26 11:37, Michael W. Lucas wrote: > Hi, > > I got my first kernel panic in ten years! (Was attempting to add a 32nd > zvol-backed drive to a bhyve guest.) > > This host didn't have kernel debug symbols. I went to install them and > discovered my host needed updates. Thoughtlessly, I grabbed the new > kernel and the new symbols. > > Oops. You can't debug a 15.0 kernel on 15.0p3. > > bucket/var/crash;kgdb -n 0 > kgdb: couldn't find a suitable kernel image > > I have a 15.0 boot environment I could boot into, if there's a way to > install the debugging symbols? Or is there a way to pull the old > kernel and debug symbols down on the current BE? > > On dch's advice I tried setting the pkg.conf URL: > url: > "pkg+https://cloudfront.aws.pkgbase.freebsd.org/FreeBSD:15:amd64/base_release_ > 0/" > but got: > pkg: No SRV record found for the repo 'FreeBSD-base' > > > Thanks, > ==ml > > PS: I shoulda just done make world, that way I'd have /boot/kernel.old. ;-) > > >
Going in somewhat of a different direction, these days you should be explicit about how the 15.0 kernel came to be in place for asking such questions: ) pkgbase (with what packages installed)? ) distribution sets (with which installed)? ) source build and install (with what built and installed)? The proper activity to get symbols in place likely depends on such. If pkgbase is involved, then there are more considerations related to ending up failing its self checks if the system is modified via other than pkg. Also pkgbase has a non-git /usr/src/ tree (if one has had it installed). I've never seen anything explicit about how to deal with debugging for pkgbase installs, much less for when symbols and/or source was not installed up front. Whatever the case for such, well identifying the installation technique when asking questions is appropriate. -- === Mark Millard marklmi at yahoo.com
