[9fans] simple venti question.

[9fans] simple venti question.

[ron minnich]

The situation:
- standard laptop with fossil+venti
- dead fossil (filled up, can't boot the machine at this point as fossil
  is full -- not sure yet why this keeps happening ...)
- I lost the venti score file I was keeping
- Plan is to boot the bootable cd, do the hget of the index,
  do the 8.printarenas thing, get a good vac, and restore fossil.

Do I take the lastest or the firstest vac that gets printed when I do the
8.printarenas ?

If I boot the CD, and I do this:
venti/conf /dev/sdC0/arenas0 > /tmp/v.conf I get this:
index main
areas /dev/sdC0/arenas0
areas /dev/sdC0/arenas1
isect /dev/sdC0/isect0
isect /dev/sdC0/isect1

which matches the machine as it has been running.

I have no venti environment variable at this point.
If I run
venti/venti -c /tmp/v.conf

I get this:

err 2: mistmatched number index sections: found 2 expected 1
fatal venti error blah blah blah

Not sure what this is, anyone seen it?

ron

Re: [9fans] simple venti question.

[ron minnich]

On Tue, 3 Feb 2004, ron minnich wrote:

> areas /dev/sdC0/arenas0

arenas, sorry.
> areas /dev/sdC0/arenas1

ditto.

Re: [9fans] simple venti question.

[Fco.J.Ballesteros]

> Do I take the lastest or the firstest vac that gets printed when I do the
> 8.printarenas ?

Regarding the vac for your fossil, perhaps you could use the
script from rsc that I include below to print the last fossil vac number, and then do
a fossil/flfmt -v using that vac to recover from venti.

> err 2: mistmatched number index sections: found 2 expected 1
> fatal venti error blah blah blah

I've never seen this. I think that it's just that your venti
index is wrong. But from the man page it seems that you'd better
try venti/checkindex and then venti/buildindex, should checkindex
fail.


          venti/checkindex [ -f ] [ -B blockcachesize ] venti.config
          tmp
         venti/buildindex [ -B blockcachesize ] [ -Z ] venti.config
          tmp
          Buildindex .... This command is typically
          used to build a new index for a Venti system when the old
          index becomes too small, or to rebuild an index after media
          failure.

            Checkindex examines the Venti index described in
          venti.config. The program detects various error conditions
          including: blocks that are not indexed, index entries for
          blocks that do not exist, and duplicate index entries.  If
          requested, an attempt can be made to fix errors that are
          found.



#!/bin/rc

rfork e

switch($#*){
case 0
	file=/dev/sdC0/fossil
case 1
	file=$1
case *
	echo 'usage: lastfossil [/dev/sdC0/fossil]' >[1=2]
	exit usage
}

fn h2d { echo 'ibase=16;' print `{echo $1|tr a-z A-Z} |bc }
header=`{dd -bs 1024 -iseek 128 -count 1 -if $file >[2]/dev/null| xd -b | sed 1q}
bsize=`{h2d $header(8)^$header(9)}
super=`{h2d 0$header(10)^$header(11)^$header(12)^$header(13)}
x=`{dd -bs $bsize -iseek $super -count 1 -if $file >[2]/dev/null | xd -b |
	sed '1,2d;3s/^.......  .. .. //;4s/^.......  (.. .. .. .. .. ..).*/\1/;4q' | tr -d ' '}
echo vac:^$x(1)^$x(2)

Re: [9fans] simple venti question.

[plan9]

> - dead fossil (filled up, can't boot the machine at this point as fossil
>   is full -- not sure yet why this keeps happening ...)

You probably haven't set up snapshot timeouts for your fossil.  From
fossilcons(8):

...
          [ fsys name ] snaptime [ -a hhmm ] [ -s interval ] [ -t
          timeout ]
...
          Snaptime displays and edits the times at which snapshots are
          automatically taken.  An archival snapshot is taken once a
          day, at hhmm, while temporary snapshots are taken at multi-
          ples of interval minutes.  Temporary snapshots are discarded
          once they are timeout minutes old.  With no arguments,
          snaptime prints the current snapshot times.  The -a and -s
          options set the archive and snapshot times.  An hhmm or
          interval of `none' can be used to disable that kind of auto-
          matic snapshot.  The -t option sets the snapshot timeout.
          If timeout is `none', temporary snapshots are not automati-
          cally discarded.  By default, all three times are set to
          `none'.
...

Without that, your fossil will fill up with as many snapshots as fit.
I believe you can make space in your fossil by using the epoch
fossilcons command.  But the proper way is to set up a timeout for it
to discard old snapshots.

When this happened to me ages ago, I was able to boot using my old kfs
partition and then start the fossil manually (it wasn't able to serve
files, but the my fscons was available), connect to its console, use
epoch to clear some space and finally boot using fossil again.

Hope this helps,
Robby

Re: [9fans] simple venti question.

[ron minnich]

On Wed, 4 Feb 2004 plan9@blueyonder.co.uk wrote:

> > - dead fossil (filled up, can't boot the machine at this point as fossil
> >   is full -- not sure yet why this keeps happening ...)
>
> You probably haven't set up snapshot timeouts for your fossil.  From
> fossilcons(8):

if only. That was not it.

ron

Re: [9fans] simple venti question.

[ron minnich]

On Wed, 4 Feb 2004 plan9@blueyonder.co.uk wrote:

> When this happened to me ages ago, I was able to boot using my old kfs
> partition and then start the fossil manually (it wasn't able to serve
> files, but the my fscons was available), connect to its console, use
> epoch to clear some space and finally boot using fossil again.
>

thanks, but it suicides now when I start it up ...

ron