Conference decwet::networker

444.0. "looking for default pool" by MLNCSC::VOCI () Thu Feb 27 1997 09:30

    Hello,
    
    I'm working with dec nsr 4.2B on Digital Unix 4.0, save versions on the
    client and on the server.
    
    I see the following behaviour:
    
    I define pool1 (with label lab1) that uses an optical disk (which is
    actualy a file on my disk since I don't have a tape). This pool is
    used only by one client that does allways a full backup, this client
    is the only one in group1. Everything is quite simple since I'm doing
    some testings for a customer problem. The level defined in the pool is 
    full.
    
    When I start group1 I see the following messages:
    
    Thu 18:12:23 savegroup info: starting group2 (with 2 clients)
    Thu 18:12:29 savegroup info: zaghi1.mln.dec.com:/usr/users/gea/nsr2 no
    cycles found in media db; doing full save
    Thu 18:12:31 media info: label a new writable volume for pool `pool2'
    Thu 18:12:32 media waiting event: backup to pool 'pool2' waiting for 1
    writable backup disk
    Thu 18:13:12 /usr/nastri/dev2 mounted optical disk label2.001
    Thu 18:13:12 media event cleared: backup to pool 'pool2' waiting for 1
    writable backup disk
    Thu 18:13:12 zaghi1.mln.dec.com:/usr/users/gea/nsr2 saving to pool
    'pool2' (label2.001)
    Thu 18:13:14 zaghi1.mln.dec.com:/usr/users/gea/nsr2 done saving to pool
    'pool2' (label2.001)
    Thu 18:13:51 media info: label a new writable volume for pool `Default'
    Thu 18:13:51 media waiting event: backup to pool 'Default' waiting for
    1 writable backup disk
    
    The problem is caused by the 'backup to pool 'Default' waiting for 1
    writable backup disk' the pool default isn't used and has no devices
    associated with it.
    
    If I add level 9 in pool1 the problem no longer appears.
    
    Is this a know feature?
    I tried to access dec west patch repository but is down.
    
    Thanks,
    Gea

	Yes the client index save is set for a level 9 on an
	incremental backup.

    But I was doing a full backup, and it was the first one on theta
    particular client.
    
    Gea 

Pool Default is not used for the client data, but for
the client's index (which is on the server), which is
saved at level 9 (always).  One can change that, but
it can slow down disaster recovery.  This is discussed
in various places in the old notes conference (DECWET::DECNSR).
You've already got the workaround that has been suggested
for others encountering this.

	Just to be clear.

	The only way to be sure that the index save will go to a full
	pool is to specify a full save.  This can be done by the
	'savegrp -l full' command or by a scheduled save that calls for
	a full via the calendar associated with that client.

	Any save you start by the group control will be an incremental.
	If there has not been a previous full for the client it will
	promote the client data saves to fulls, this will still
	leave the index/bootstrap save at level 9.

    Thanks for the infos.
    
    Gea