[Gluster-users] upgrade path 2.0.0rc1 -> 2.0.9 ?

[Daniel Maher]

Hello all,

I have a small gluster setup with two servers and two clients doing 
client-side replication to the servers.  All of the nodes are currently 
running 2.0.0rc1 .

In preparation for an upgrade to 2.0.9 i set up a series of virtual 
machines with the same OS and environment as the actual machines.  After 
compiling and installing the new binaries, i restarted the various 
services and mounts, and everything hummed along nicely.

Unfortunately, this was not the case for my actual machines. :(  I 
upgraded one of the servers, then restarted the gluster service, and saw 
nothing but this over and over in the logs (at high speed) :

[2010-01-26 11:09:26] N [server-protocol.c:7825:notify] server: 
10.0.0.16:853 disconnected

10.0.0.16 being, in this case, one of the clients.  On the same client, 
the log was full of these messages :

2010-01-26 11:18:13 E [client-protocol.c:5604:client_setvolume_cbk] 
nasdash-01: SETVOLUME on remote-host failed: UUID not specified

nasdash-01 being, in this case, the server noted above.

I thus have two questions :

1. What does the « SETVOLUME on remote-host failed: UUID not specified » 
error actually mean ?  I realise that it means i need to specify the 
UUID - what is that, why do i need to specify it, and how do i go about 
that ?  (A search for « UUID » on the gluster wiki produces zero hits.)

2. What is the Official Gluster Approved Upgrade Methodology for going 
from 2.0.0rc1 to 2.0.9 ?

Thank you.

(p.s. We're not ready to move to the 3.0.x branch...)


-- 
Daniel Maher <dma+gluster AT witbe DOT net>