OpenBSD Journal

April 20th snapshot currently broken:

Contributed by sean on from the look before you leap dept.

As may or may not have been seen on misc:
From: Ted Unangst 
Subject: don't use today's snapshot

the snapshot currently on the ftp servers is broken.  depending on your
system, it will die during boot or hang when the network comes up.  the
code has been fixed in cvs, but that doesn't fix the snapshot.  i can't
say when a new one comes out, but don't use one dated apr 20.

if you have installed the snapshot, and cannot build a new kernel for some
reason, the following will enable you to boot. at boot> prompt, type -d 
to enter debugger.  then type "w uvm_km_pages_lowat 128", and then c to 

(Comments are closed)

  1. By Xenotrope ( on

    It's been my experience that the snapshots have been broken since April 15, 2004, at least for the i386 architecture. I haven't gotten one yet to install successfully. I managed to find an i386 snapshot from one of the mirrors dated April 14, so I've been using that one in the meantime. I find the appearance of these recent snapshots very odd: new snapshots usually stop showing up for a month prior to a release. Is this just a New Way of Doing Things?

    1. By Chris Cappuccio ( on

      No, it's not a "New Way", it is the way it has always been. Snapshots are just that; a compilation of the current DEVELOPMENT source tree!!! Some times, there is a mistake in the development... I think the fact that everyone expects the snapshots to be production quality (and that the general rules for committers help to maintain this level of quality) speaks for itself. (Hell, who else ever apologises when their development tree has a show-stopper? :)

      1. By bob ( on

        snaps are for testing imho, best bob btw know one when the 3.5 is realy shipping? (for the long term pre-order folks.)

    2. By Otto Moerbeek ( on

      1. If we do not get a report, we are not able to fix your problem. I would suggest to try the most recent snapshot (it should work) and post on misc@ (including dmesg, other details, etc) if it doesn't.

      2. Thaw has set in. The tree is no longer locked. We are working on 3.5-current.

    3. By Peter Hessler ( on

      the 15th installed just fine for me (once I used a switch port that was in "non-broken" mode)

  2. By mnology ( on

    Figures. I put off upgrading from 3.3 until yesterday, and subsequently hosed things. I realize my problems are still mostly due to my 3.3 boot loader and the a.out change. After manually downloading the packages( workaround for the network issues ) and starting the install from a mounted disk, the update script had issues deleting some perl/ssl directories.

  3. By tedu ( on

    as i noted later, the 20th snap may be ok, it was the 19th that was broken, but this could vary by arch. anyway, install next week and all will be well. the main point was, yeah, we know there were problems, and they're fixed already, so a dozen more bug reports won't help much. :)

  4. By rogerwilco ( on

    I used the 22th snap on my G4 mac, and it seems to be working fine....... I guess this is all i386 issues... :P


Copyright © - Daniel Hartmeier. All rights reserved. Articles and comments are copyright their respective authors, submission implies license to publish on this web site. Contents of the archive prior to as well as images and HTML templates were copied from the fabulous original with Jose's and Jim's kind permission. This journal runs as CGI with httpd(8) on OpenBSD, the source code is BSD licensed. undeadly \Un*dead"ly\, a. Not subject to death; immortal. [Obs.]