First my thanks to all those who responded.
Many thanks to
Arthur Darren Dunham
Viet Q Hoang
Only option 2 (rsh dd) is feasible.
1. UFSDUMP requires a local file system and NFS filesystems could not be dumped.
(Somehow I missed this point).
2. Samba backup is not worthwhile.
3. Go for some backup management software.
1. Instead of rsh dd, go for rmt or tapehost:/dev/rmt/xx
2. Increase blocking factor upto 512 blocks.
3. Going for private network (like SAN). Too costly.
4. Upgrade tapedrive to some high performance tapedrives like DLT.
5. Use buffer command for performance. Start ufsdump remotely from tapeserver.
(I haven't tested this.)
I apologize that I haven't furnished much info earlier. Earlier I was getting a
throughput of about 750K for a 2gig filesystem.
I tested backup with 2 changes. Instead of rsh dd, I used tapehost:/dev/rmt/xx
and changed the blocking factor. I got the best performance for default blocking
factor!! i.e. blocking factor of 32K consistantly gave me over 950K performance
touching 1Mb occasionally. Blocking factors of 126,256 and 512 usually gave
about 830-870K. (This is a rough benchmark... don't read much into it).
I will go for options 4 & 5 sometime later.
Thanks again for your replies.
Here is my original post
15/09/2000 06:09:32 PM
Subject: Remote Backup Performance Issues
Currently we are having single tape drive (HP DDS3) and we are using that to
take remote backup on multiple sun machines. But the backup process is
Ours is a 100Mbps switched network environment and all the machines we take
backup are in the same subnet. In that case, I would like to know which method
will be efficient.
1. Mounting the exported file system on the machine that is connected to
tapedrive and take ufsdump.
2. Output ufsdump locally to stdout and executing rsh dd on the remote
3. All solaris machines are running samba. Mapping them as network drive
and taking NT Backup. (But I wouldn't prefer this because recovery will be
difficult. But this can be useful for user files. )
4. Any other method.
I would like to mention the following points
1. CPU Load is not heavy. (30 - 40% on working hours)
2. I prefer ufsdump only because recovery is quicker.
It would be better if you could differentiate each method with the performance
gain in terms of time. Also what blocking factors will give better performance
(I am using 126blocks ..default for cart) and anyother tips that will improve
Thanks in advance. I will summarise.
U BEFORE POSTING please READ the FAQ located at
. and the list POLICY statement located at
A To submit questions/summaries to this list send your email message to:
A To unsubscribe from this list please send an email message to:
E and in the BODY type:
R unsubscribe sun-managers
. unsubscribe sun-managers firstname.lastname@example.org
L To view an archive of this list please visit:
This archive was generated by hypermail 2.1.2 : Fri Sep 28 2001 - 23:14:18 CDT