[332] in Athena_Backup_System
ABS vldb syncing question
daemon@ATHENA.MIT.EDU (ted@MIT.EDU)
Wed Oct 23 11:16:02 1996
From: ted@MIT.EDU
To: delgado@MIT.EDU
Cc: athena-backup@MIT.EDU
Date: Wed, 23 Oct 1996 11:15:05 EDT
Diane,
I just realized a potential problem in the method that the ABS master
updates it's idea of where the volumes are. If I understand
correctly, when the master updates it's volume location info, first it
sets the server to 'UNKNOWN' for the volumes it's curious about, then
it calls VL_ListByAttributes, and then it sets the server location
depending on what it finds.
The problem that I see with this is that if the VL_ListByAttributes
call fails for any reason, then the master has lost it's idea of where
the volumes are. Am I correct that this problem exists?
If so, is there any reason the algorithm can't be changed to set the
server to 'UNKNOWN' only after we know that the VL_ListByAttributes
call is successful? If this change can't be made, will we be able to
easily diagnose this problem if/when it happens and will there be an
easy way to fix the master's volume location info?
--Ted
