Summary:::Solstice backup 5.5.1 problem on Solaris 2.51

From: Raj, Dilip <>
Date: Mon May 21 2001 - 15:04:03 EDT

I am Extremely grateful to those that responded:

Michael Auria
Julian John C
Darren Durham
Stuart Whitby
Patrico Mora

The question was in two portions:

Original Questions:

Q:1.	How would I ensure that when I start my scheduled backup, it will
always start from the beginning of tape. I basically want to over-write next
month's tape. ( For example, last month's 17th Tape will be recycle, and
start from the beginning when I use it for this month 17th). I tried using
the device called  ( /dev/rmt/0m   but still was not able to write from the


I still did not got concrete answer for this one. Someone suggest to use
/dev/rmt/0m, it still didn't work.
Someone suggested to use retension policy. Also, someone suggests to use the
auto-recycle mode, it still no result.
I will call Sun, and then summerize it

Here are some suggestions:

1.	Ensure that the tape is past its retention policy, and
that no further incremental backups depend on that 
2.	If a tape is full and all data on it is recyclable it will be
written from
the beginning overwritting the entire tape.

Assume a full backup takes 5 1/2 tapes. I make sure that my juke box has at
least 6 full tapes that nwadmin thinks are recyclable. 

In nwadmin, click "volumes". All the tapes that have "recyc" in the mode
column and "full" in the %Used column will be overwritten.

3.	Tapes are always appended to. Only if the tape is full and the tape
recyclable will it be overwritten.

You can force a tape to have zero data by relabeling the tape with its
current label.

Q:2.	My three clients are working fine, but when I tried to restore some
of the files from one clients ( monroe) , I am getting following errors: 
	From Clients:
	#  nwrecover &
	 nwrecover: RPC error, Remote address unknown: Program not

This was definitely a RPC services that were used by this software.
I had commented out some of the RPC services. I tried one by one, and one
that had fixed was


I turned on below services, in inetd.conf file
# rpc.cmsd is a data base daemon which manages calendar data backed
# by files in /var/spool/calendar
100068/2-4      dgram   rpc/udp wait root /usr/openwin/bin/rpc.cmsd

After this I was able to resotre from my clients.

Here are the suggestions:

1) verify connectivity and hostname lookups (ping server)
2) name the server for recovery (nwrecover -s server)

Stuart Whitby had a best suggestions also,

Sounds like an RPC issue, though that's not necessarily
the root cause.  Try launching from the server and running
a directed recover, try just running "recover" from the
command line (which is far nicer for a unix admin anyway),
check that forward and reverse name resolution (checking 
by name and IP) give the same answer and nsure that 
rpcbind is running.  Check for more 
info on this.

Dilip Raj
System Administrator.
Access Control Center.
Motorola Broadband Communication Sector.
Phone: (858)-404-3878
Received on Mon May 21 20:04:03 2001

This archive was generated by hypermail 2.1.8 : Wed Mar 23 2016 - 16:24:55 EDT