Coda File System

Re: Adding non-SCM/reinstalling data on crashed non-SCM/SCM

From: Jan Harkes <jaharkes_at_cs.cmu.edu>
Date: Sun, 20 Jul 2003 13:35:41 -0400
On Sun, Jul 20, 2003 at 02:30:35AM -0400, UpFront Technology wrote:
> The volumename of a replica is the replicated volume name + an uniquifying
> number, i.e. rootvolume.0.
> 
>     volutil create_rep <partition> <replicaname> <replicated-id>
> <replica-id>
> 
> e.g.
> 
>     volutil create_rep /vicepa volume.1 7F000010 2000007 "   [2000-07-24
> 13:56:06]  .
> ----------------------
> 
> Is this to be perfomed on the non-SCM?  I assume yes.

Correct, if you want to run this from another server (such as the SCM)
you would have to do something like,

    SCM$ volutil -h non-SCM create_rep /vicepa volumename 7F000XXX replicaid

> I then believe that the next step is to run:   > bldvldb.sh    on the SCM,
> this I assume replicates the RVM data (which is what is used to rebuild
> data, not a copy of /vicepa) to the non-SCM.

This repopulated the volume location database (VLDB), every server has a
copy of this file and uses it when a client asks for the location of a
specific volume replica.

> 1. recreate the volumes with appropriate id's on the non-SCM
> 2. run > bldvldb.sh on the SCM?

You can also run 'bldvldb.sh non-SCM', then it will only fetch the
current volumelist from the newly initialized server 'non-SCM'. Ofcourse
the newly created volume replica will be completely empty. So the next
step would be to use a Coda client to trigger resolution for each file.
This is not too hard as long as you know where your volumes are mounted
in the tree.

   cd /coda/path/to/volume ; ls -lR

Jan
Received on 2003-07-20 13:38:45