Coda File System

Re: Replication server problems

From: Achim Stumpf <newgrp_at_gmx.de>
Date: Mon, 22 Jan 2007 09:55:23 +0100
Jan Harkes wrote:
> That's a little strange, I thought you had 2 volumes, the root volume
> and a triply replicated testvol.
>   
No, the testvol was in the first new installation, because I wasn't sure 
what's going on, I decided to reinstall the whole thing without a testvol.
> Ok, let's check if the volume location database is out of sync,
>
>     # bldvldb.sh clusty1.mytest.de
>
> This will fetch an up-to-date list of volumes from clusty1 and rebuild
> the VLDB file. If clusty1 really doesn't have it's replica, then the
> createvol should succeed.
>   
[root_at_clusty1 db]# bldvldb.sh clusty1.mytest.de
Fetching volume lists from servers:
V_BindToServer: binding to host clusty1.mytest.de
GetVolumeList finished successfully
 clusty1.mytest.de - success
V_BindToServer: binding to host clusty1
VLDB completed.
[root_at_clusty1 db]# createvol_rep / clusty1.mytest.de clusty2.mytest.de 
clusty3.mytest.de
/ already exists as a replica on clusty1.mytest.de

> If it still fails, then clusty1 still thinks it has that volume replica,
> we can remove it manually, We first need to know the internal id of the
> volume replica and then we can tell the server to drop the volume.
>
>     # volutil -h clusty1.mytest.de info /.0 | grep volume
>     Volume header for volume 01000001 (/.0)
>
>     # volutil -h clusty1.mytest.de purge 0x01000001 /.0
>     # bldvldb.sh clusty1.mytest.de
>
> At this point createvol_rep should work.
>
> Jan
>   
[root_at_clusty1 db]# volutil -h clusty1.mytest.de info /.0 | grep volume
V_BindToServer: binding to host clusty1.mytest.de
Volume header for volume 01000001 (/.0)
[root_at_clusty1 db]# volutil -h clusty1.mytest.de purge 0x01000001 /.0
V_BindToServer: binding to host clusty1.mytest.de
Volume 01000001 (/.0) successfully purged
[root_at_clusty1 db]# bldvldb.sh clusty1.mytest.de
Fetching volume lists from servers:
V_BindToServer: binding to host clusty1.mytest.de
GetVolumeList finished successfully
 clusty1.mytest.de - success
V_BindToServer: binding to host clusty1
VLDB completed.
[root_at_clusty1 db]# createvol_rep / clusty1.mytest.de clusty2.mytest.de 
clusty3.mytest.de
Replicated volumeid is 7f000001
creating volume /.0 on clusty1.mytest.de (partition /vicepa)
V_BindToServer: binding to host clusty1.mytest.de
creating volume /.1 on clusty2.mytest.de (partition /vicepa)
V_BindToServer: binding to host clusty2.mytest.de
creating volume /.2 on clusty3.mytest.de (partition /vicepa)
V_BindToServer: binding to host clusty3.mytest.de
Fetching volume lists from servers:
V_BindToServer: binding to host clusty3.mytest.de
GetVolumeList finished successfully
 clusty3.mytest.de - success
V_BindToServer: binding to host clusty2.mytest.de
GetVolumeList finished successfully
 clusty2.mytest.de - success
V_BindToServer: binding to host clusty1.mytest.de
GetVolumeList finished successfully
 clusty1.mytest.de - success
V_BindToServer: binding to host clusty1
VLDB completed.
<echo / 7f000001 3 01000002 02000001 03000001 0 0 0 0 0 >> 
/vice/db/VRList.new>
V_BindToServer: binding to host clusty1
VRDB completed.
Do you wish this volume to be Backed Up (y/n)? [n] n

On the client clusty4:
[root_at_clusty4 ~]# cfs whereis /coda/mytest.de/
  clusty1.mytest.de  clusty2.mytest.de  clusty3.mytest.de

So it looks good. Thanks a lot...

Now I have to figure out how to put files on the replica and then I will 
shutdown some servers and see how coda acts...


Achim
Received on 2007-01-22 04:00:47