Hello,
As I understood, the
SCM server is just different from the replicate server
in the way it is
responsible for the user database and the volume related
functionalities.
Normally, if the SCM
shutdown (due to hardware or network failure), the coda system (client and
replicate server) should
still
running...
In a configuration
with one volume replicated on two servers, the configuration is
:
# cat
/vice/db/VSGDB
E0000100 servint1
E0000102 proxytest
E0000103 servint1 proxytest
# cat /vice/db/VRList
vROOT 7f00000b 1 100000c 0 0 0 0 0 0 0 E0000100
vSHARE 7f00000c 2 100000d 2000004 0 0 0 0 0 0 E0000103
E0000100 servint1
E0000102 proxytest
E0000103 servint1 proxytest
# cat /vice/db/VRList
vROOT 7f00000b 1 100000c 0 0 0 0 0 0 0 E0000100
vSHARE 7f00000c 2 100000d 2000004 0 0 0 0 0 0 E0000103
vROOT is the root
volume wich will be automaticaly mounted
as /coda/my-realm by
the client software
THE QUESTION IS :
what happens during an SCM failure ?
The replicated
server can still properly work as there is no root volume anymore
?
I ask because I ran
many inconsistence problems while playing with Coda...
In particular when
shutting down the SCM.
I need to
understand, please help.
Marc
Cassuto.
