Coda File System

Re: Fw: creating rep_volumes without success

From: Jan Harkes <jaharkes_at_cs.cmu.edu>
Date: Thu, 7 Mar 2002 13:55:31 -0500
On Thu, Mar 07, 2002 at 05:26:32PM -0000, Mohamed Alwakeel wrote:
> o.k. VRList.new already exists and so does the directory, now I get a
> different error, here it is with and without sh -x debug:-
> Also below that is SrvLog dump
...
> [root_at_mezzo1 db]# createvol_rep codaroot2 E0000104
> V_BindToServer: binding to host mezzo1
> Servers are ()
> Specify 1 - 8 servers
> [root_at_mezzo1 db]# sh -x createvol_rep codaroot2 E0000104
...
> ++ cat /vice/db/scm
> + volutil -h mezzo1 dumpvrdb /vice/db/VRList.new
> V_BindToServer: binding to host mezzo1
> VolDumpVRDB failed with Unknown RPC2 return code 200

Ok, the failed run of createvol_rep has ofcourse left a new VRList.new
file laying around which then in turn breaks the run with the debugging
flags. I'm adding a rm -f before the call to volutil to fix that issue.

But the problem in the first run is pretty clear,

    # Derive the server list from the <vsgaddr> parameter
    SERVERS=`awk '$1 ~ /^'$VSGADDR'$/ {print $2, $3, $4, $5, $6, $7, $8, $9}' ${vicedir}/db/VSGDB`
    echo Servers are \($SERVERS\)

I.e. we look in the file /vice/db/VSGDB for any line that starts with
E0000104 to get the list of servers we need to use.

However, this line isn't found, so we bail out.

> 15:32:11 VLDB_Lookup: no more records in VLDB
> 15:32:12 VLDB_Lookup: no more records in VLDB

Hmm, I wonder why there are stray lookups for non-existing volumes,
perhaps there is some bad data lingering in the VRList/BigVolumelist,
but that should hopefully go away as soon as you managed to create the
volume.

> 15:35:21 S_VolDumpVRDB: unable to open file /vice/db/VRList.new

These are the errors associated with the volutil dumpvrdb failures ;)

Jan
Received on 2002-03-07 13:55:55