[OpenAFS] pts: no quorum elected... lowest IP says it's a clone

[Dave Broudy]

Ken Hornstein wrote:

>If I remember the "clone" logic, basically that means a clone never gets
>voted for.  That's assuming that everyone knows that it's a clone.
>  
>
Which clearly isn't the case, as the other machine keep voting for it, 
and I don't see it telling them differently when I sniff for traffic 
(assuming this would all be happening on 7002).

>Dumb question: do all your CellServDB entries look the same?
>
Possibly dumber answer: I'm using AFSDB records, and those resolve the 
same on all machines:

broudy.net              AFSDB   1 noah.broudy.net
broudy.net              AFSDB   1 minerva.broudy.net
broudy.net              AFSDB   1 goliath.broudy.net

Hmm... is that only for clients?

>  I'm
>talking about the CellServDB in /usr/afs/etc (I think that's the
>right directory).  And what does udebug report on those three
>machines on port 7003?
>  
>

I just added the server to the CellServDB (it's /etc/openafs/server 
because I'm running debian packages), and the lowest IP was just elected 
the sync site... so it works. So, I guess I have to maintain the 
CellServDB file and just use AFSDB records for the clients, where they 
seem to work. I'll try to break it for a few days and see what happens.

With empty CellServDB (just the cell name) files, vlserver worked ok, 
and elected new sync sites with one was down. Here was the 7003 output 
before I made the change:

(500) dave:louie:~$ udebug goliath 7003
Host's addresses are: 63.228.79.9 10.222.3.5
Host's 10.222.3.5 time is Sun Dec 18 23:19:13 2005
Local time is Sun Dec 18 23:19:15 2005 (time differential 2 secs)
Last yes vote for 63.228.79.9 was 0 secs ago (sync site);
Last vote started 0 secs ago (at Sun Dec 18 23:19:15 2005)
Local db version is 1134903942.61
I am sync site until 59 secs from now (at Sun Dec 18 23:20:14 2005) (2 
servers)
Recovery state 1f
Sync site's db version is 1134903942.61
0 locked pages, 0 of them for write
Last time a new db version was labelled was:
         69211 secs ago (at Sun Dec 18 04:05:44 2005)

Server (63.228.79.12): (db 1134903942.61)    is only a clone!
    last vote rcvd 0 secs ago (at Sun Dec 18 23:19:15 2005),
    last beacon sent 0 secs ago (at Sun Dec 18 23:19:15 2005), last vote 
was yes
    dbcurrent=1, up=1 beaconSince=1

Server (206.123.106.135): (db 1134903942.61)
    last vote rcvd 1 secs ago (at Sun Dec 18 23:19:14 2005),
    last beacon sent 0 secs ago (at Sun Dec 18 23:19:15 2005), last vote 
was yes
    dbcurrent=1, up=1 beaconSince=1

(501) dave:louie:~$ udebug minerva 7003
Host's addresses are: 63.228.79.12
Host's 63.228.79.12 time is Sun Dec 18 23:19:29 2005
Local time is Sun Dec 18 23:19:32 2005 (time differential 3 secs)
Last yes vote for 63.228.79.9 was 3 secs ago (sync site);
Last vote started 1 secs ago (at Sun Dec 18 23:19:31 2005)
Local db version is 1134903942.61
I am not sync site
Lowest host 63.228.79.9 was set 3 secs ago
Sync host 63.228.79.9 was set 3 secs ago
Sync site's db version is 1134903942.61
0 locked pages, 0 of them for write

(502) dave:louie:~$ udebug noah 7003
Host's addresses are: 206.123.106.135
Host's 206.123.106.135 time is Sun Dec 18 23:19:36 2005
Local time is Sun Dec 18 23:19:39 2005 (time differential 3 secs)
Last yes vote for 63.228.79.9 was 8 secs ago (sync site);
Last vote started 8 secs ago (at Sun Dec 18 23:19:31 2005)
Local db version is 1134903942.61
I am not sync site
Lowest host 63.228.79.9 was set 8 secs ago
Sync host 63.228.79.9 was set 8 secs ago
Sync site's db version is 1134903942.61
0 locked pages, 0 of them for write


Thanks,
Dave

-- 
Dave Broudy
dave@broudy.net
http://www.broudy.net/
Phone: 303.278.0908      Mobile: 703.401.5955        Fax: 303.674.6840
AIM/YIM: dbroudy         Jabber: dbroudy@jabber.org