Coda File System

Re: moving a dying server

From: Greg Troxel <gdt_at_ir.bbn.com>
Date: 16 Jun 2005 15:44:11 -0400
I've successfully moved my server, and learned a few things.

I rsynced /vice and /vicepa (really to /n0/CODA/vice{,pa} and then
symlinked).  I'm using code from December 2004, compiled on NetBSD 2,
and running on 2.99.15.  The server data was from 1.6.2, and it works
ok with the new coda and kernel bits.
 
points of note:

* (Jan just said..) On NetBSD, venus uses inode numbers for cache
  files, so you can't rsync venus state to another machine.  This is
  too bad, and violates the Principle of Least Astonishment.  It might
  be nice to cache the inodes in regular RAM, looking them up the
  first time they are used.   But this isn't a big deal.

  note to self: fix kernel bug when handed wacky inode #

* On NetBSD, the RVM LOG seems to have to be a raw device, not a file
  (hence use vnconfig).  DATA can be a file.  On starting codasrv or
  norton with a file for LOG, I got rvm log initialization errors.

* I had used dd or rsync on data and log, and the resulting log file
  was borked.  Obviously you should shutdown the server first, and I
  probably didn't.

* norton-reinit was able to dump, but I tried to load and it chewed up
  minutes of cpu time on a PIII-933.  I gave up and regrabbed the LOG
  file from my old server.  I regrabbed most of the DATA file before
  it crashed, and it compared the same (eof on file DATA.new).   I was
  then able to start up codasrv, and it seems ok.

I still don't know what's wrong with my old hardware, but I'm 99.5%
sure I got the coda server moved without data loss, which is nice
(that's all the machine did anyway).

-- 
        Greg Troxel <gdt_at_ir.bbn.com>
Received on 2005-06-16 15:48:30