Re: salvaging a previous checkpoint

• Date: Wed, 28 Feb 1996 15:14:02 PST
• From:
  Pavel Curtis <pavel@parc.xerox.com>
• Cc: moo-cows@parc.xerox.com, adam.jenkins@flinders.edu.au
• In-reply-to: Kim Silk-Copeland's message of "Tue, 27 Feb 1996 22:29:59 PST" <199602280621.BAA29398@gold.interlog.com>

Kim Silk-Copeland writes:
> and now I get a login error when I try to connect, and I disconnect with
> the welcome screen still there. The current dump.db contains the error,
> and the dump#xxx#.db also contains the error. My last unspoiled version
> of the db is from weeks ago, and much work is lost if I revert to that 
> one. Any hints??

You should shut down the running MOO (by typing `kill -INT XXX', where XXX is
the UNIX process number of the main MOO process); that will write out the
(buggy) database one last time.  Then try installing 1.8.0beta3 (that is, the
most recent test version of 1.8.0) and running it with the following command
line:

	moo -e dump.db dump.db.new

The server will come up in `Emergency Wizard Mode', allowing you to type
`program' and `list' commands and to evaluate MOO statments and expressions,
all from your UNIX shell.  You should use `list' and `program' to fix the login
verb and then use `quit' or `continue' to end the emergency mode (type `help'
to see what the difference is between these two).

You will now be stuck, I'm afraid, using 1.8.0beta3, since it writes database
files that can't be read by older versions of the server.  This is probably OK,
though, since 1.8.0beta3 seems to be reasonably stable and will soon be
replaced by the production release of 1.8.0 (which *can* read beta3's DB
files).

This being the case, you should carefully read through the ChangeLog.txt
entries for all of the releases since the one you used to be running, looking
for the string `NOTE'; it marks further changes you should make to your DB
before considering the upgrade complete.

> Also, how to avoid this in the future? Do you periodically shutdown the
> moo server, store the dump.db in a safe place, then restart??

I have the sysadmins around here back up the relevant file system every night
onto tape and save those tapes for several weeks, so that I can always get back
fairly recent data in case of problems.

	Pavel

• salvaging a previous checkpoint
• From: Kim Silk-Copeland <ksc@interlog.com>

• Prev: Re: salvaging a previous checkpoint
• Next: Re: salvaging a previous checkpoint
• Index: MOO-cows
• Thread: MOO-cows