[OpenAFS] adding a database server to a cell

[Blake Atkins]

I run into this all the time with my cells. Every time I add a database 
machine to the cell, I have to 'bos restart <server> -all' on every single 
database server to get it to work properly. I do not normally get a core 
file, though.



On Thursday 11 October 2001 12:29 pm, you wrote:
> I'm stumped, I've checked hostnames ... at first go I had tried
> making this work on hosts that happened to be multi-homed - the
> FAQ says that as of AFS 3.4 database servers may not be multi-homed
> is this still true of openafs?
>
> Anyhow, it still fails on working from a single-interface
> system, here's the order, taken from the AFS quick-beginnings
> doc file:
>
> the operating AFS database server is a.fwsystems.com
>
> made copies of /usr/afs/etc/* on the new server (b)
> started bosserver
>
>
> bos create b.fwsystems.com upclientetc simple "/usr/afs/bin/upclient
> a.fwsystems.com /usr/afs/etc" -cell athena.fwsystems.com
>
> bos addhost a b.fwsystems.com
>
> verified that the new server was receiving updatated etc files
>
> bos listh b
> Cell name is athena.fwsystems.com
>     Host 1 is a.fwsystems.com
>     Host 2 is b.fwsystems.com
>
>
> bos create b.fwsystems.com buserver simple /usr/afs/bin/buserver
> gives:
> /usr/afs/bin/buserver: problems with host name Ubik init failed
> /usr/afs/bin/buserver: problems with host name Ubik init failed
> /usr/afs/bin/buserver: problems with host name Ubik init failed
> /usr/afs/bin/buserver: problems with host name Ubik init failed
> .... 13 total messages
>
> On the console of the primary database server (a):
>
> x.y.z.13
>
> x.y.z.13
>
> x.y.z.13
>
> x.y.z.13
> .... 13 total messages
>
> continue on b:
>
> bos create b.fwsystems.com ptserver simple /usr/afs/bin/ptserver
> bos create b.fwsystems.com vlserver simple /usr/afs/bin/vlserver
>
> (I'm not running kaserver, all authentications are via aklog to
> the K5 kdc ... if kaserver were running, that would be the
> first step here ... is this perhaps the problem????)l
>
> on a, and b:
>
> do 'bos restart host'
>
> And the 'correct' protection and volume location servers fail on the
> primary
>
> bos status a
> Instance buserver, has core file, currently running normally.
> Instance ptserver, temporarily disabled, stopped for too many errors,
> currently shutdown. Instance vlserver, temporarily disabled, stopped for
> too many errors, currently shutdown. Instance fs, has core file, currently
> running normally.
>     Auxiliary status is: file server running.
> Instance upserver, has core file, currently running normally.
>
> The relavent logs
>
> PtLog:
>
> ptserver: problems with host name Ubik init failed
> mary address
> Thu Oct 11 11:42:42 2001 Inconsistent Cell Info from server: Thu Oct 11
> 11:42:42 2001 Local CellServDB: x.y.z.2
> Server 1: x.y.z.13
> Thu Oct 11 11:42:42 2001 Inconsistent Cell Info on server: x.y.z.13
>
> and VLLog:
>
> Thu Oct 11 11:42:43 2001 Using x.y.z.2 as my primary address
> Thu Oct 11 11:42:43 2001 Inconsistent Cell Info from server: Thu Oct 11
> 11:42:43 2001 Local CellServDB: x.y.z.2
> Server 1: x.y.z.13
> Thu Oct 11 11:42:43 2001 Inconsistent Cell Info on server: x.y.z.13
> vlserver: Ubik init failed with code 5385
>
>
> So the new server is labelling itself 'Server 1:' (Pt&VLLogs above)  ?????
>
> I'm sure I missed something in an early step, Looking at the one populated
> database on the new server (bdb.DB0), it has learned the names of all of
> the volume sets on the two fileservers. - so it's seems to be getting the
> backup database ok, the prdb* and vldb* are essentially empty.
>
> Also no /usr/afs/local/sysid file has been created?
>
> pointers on what I've overlooked?
>
> thanks,           forrest
> _______________________________________________
> OpenAFS-info mailing list
> OpenAFS-info@openafs.org
> https://lists.openafs.org/mailman/listinfo/openafs-info