From Bugzilla Helper: User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.6) Gecko/20040115 Description of problem: I'm trying to install FC devel onto one of my machines. After I click "next" on the "installation will start once you click Next" screen (after you select the packages, etc.) it format the root partition (I have no other partitions to format in my case), then produces this traceback with the phrase "error: match tag mismatch". (I will attach an anacdump.txt file to this bug) Version-Release number of selected component (if applicable): anaconda-9.91-0.20040226192747, rpmdb-fedora-1.90-0.20040228 How reproducible: Always Steps to Reproduce: 1. Boot from boot.iso. Optionally specify other filesystems or vnc (those options made no difference for me either way). 2. Begin installation over HTTP or NFS. (Other methods may work, but I didn't test those.) 3. Choose a server install. (Other types of installs may work, but I didn't test those.) Package selection doesn't seem to matter as far as I can tell, but "Minimal" is what I used most during testing. (I did not try "Everything" however.) Actual Results: Formats filesystem, then produces traceback Expected Results: Formats filesystem and proceeds uneventfully with installation Additional info: With the previous day's devel (same version of anaconda, but rpmdb-fedora-1.90-0.20040227) I do not experience this bug. If this happens to be due to a corrupted or inconsistent installation source, or anything of that sort, then a more user-friendly error message would be really nice.
Created attachment 98130 [details] anacdump: NFS install, Minimal package set, ext3 root fs I have two other anacdumps (one for HTTP install, onto an XFS filesystem, with a customized package set; the other is HTTP install, onto an XFS filesystem, with Minimal package set). I can attach one or both if you're interested.
By the way, I forgot to mention, boot.iso for 20040227 and 20040228 was bit-for-bit identical (at least on my downloads). Just in case this matters...
This is usually random tree horkage. If you're not seeing with current trees, closing.
I haven't tried current trees; my employer is considering deploying Red Hat Enterprise Linux and I recently noticed that I somehow have 2 RHEL subscriptions that will expire at the end of the month, so I've been playing with that instead. However, I think this problem disappeared with the trees from shortly after I filed this bug, so I'll leave the bug closed. (It would be nice if the error message could be nicer/easier to understand, but oh well.)