These registry settings are causing backup speed degradation in my 
environment, and I am curious if anyone could explain why.

I have a HP StorageWorks DL380 g5, running Storage Server 2003 x64 with 
dual 2.3 Quad Core processors, and 5 GB RAM.

I have a 2TB SAS connected array connected to a HP P800 controller, and 
then a HP Ultrium 920 SAS LTO-3 drive connected to a different 
controller (SAS LSI 3000)

Using the original NTBackup registry values for the keys below 
(32,512,9), I was able to backup a 36GB file in 7 Min 13 seconds.  I 
implemented these performance registry keys below (64,1024,16), ran the 
backup job again, and it took 11 min 47 seconds.

What could be causing these "performance" changes to impact backup 
speeds so greatly?

Klint

Michael B. Smith wrote:
>
>
>       *Registry Changes for Optimizing NTBackup*
>
> The first optimization to make for NTBackup performance is to change 
> some registry keys that affect buffering. These changes can have a 
> very positive impact on performance when writing to tape, and a 
> smaller impact when writing to disk. They are as follows (in batch 
> file syntax):
>
>  
>
> reg add "HKCU\Software\Microsoft\Ntbackup\Backup Engine"
>         /v "Logical Disk Buffer Size" /t REG_SZ /d 64 /f
>
> reg add "HKCU\Software\Microsoft\Ntbackup\Backup Engine"
>         /v "Max Buffer Size" /t REG_SZ /d 1024 /f
>
> reg add "HKCU\Software\Microsoft\Ntbackup\Backup Engine"
>         /v "Max Num Tape Buffers" /t REG_SZ /d 16 /f
>
>  
>
> These registry changes double the default values. Do note that they 
> affect HKEY_CURRENT_USER, and not HKEY_LOCAL_MACHINE as you might 
> expect. Therefore, you should execute NTBackup under the desired user 
> to create the registry key before you attempt to set the above 
> registry values.
>
>  
>
>  
>
> Regards,
>
>  
>
> Michael B. Smith, MCITP:SA,EMA/MCSE/Exchange MVP
>
> My blog: http://TheEssentialExchange.com/blogs/michael
>
> Link with me at: http://www.linkedin.com/in/theessentialexchange
>
>  
>
> *From:* Joseph L. Casale [mailto:[EMAIL PROTECTED]
> *Sent:* Tuesday, October 07, 2008 11:59 AM
> *To:* NT System Admin Issues
> *Subject:* RE: Comparing NT Backup speeds
>
>  
>
> Michael,
> What are these tweaks you speak of?
> jlc
>
>  
>
> *From:* Michael B. Smith [mailto:[EMAIL PROTECTED]
> *Sent:* Tuesday, October 07, 2008 6:26 AM
> *To:* NT System Admin Issues
> *Subject:* RE: Comparing NT Backup speeds
>
>  
>
> Well a local device will usually be faster than a remote device.
>
>  
>
> Ntbackup, with the registry tweaks, gives me about 1 GB per minute 
> locally. But I don't have a dat-72 to compare to. My home GB LAN with 
> a cheap crappy switch copies about 50 MB/min. So I'm thinking that two 
> hours seems more likely than 9 hours.
>
>  
>
> Regards,
>
>  
>
> Michael B. Smith, MCITP:SA,EMA/MCSE/Exchange MVP
>
> My blog: http://TheEssentialExchange.com/blogs/michael
>
> Link with me at: http://www.linkedin.com/in/theessentialexchange
>
>  
>
> *From:* HELP_PC [mailto:[EMAIL PROTECTED]
> *Sent:* Tuesday, October 07, 2008 7:28 AM
> *To:* NT System Admin Issues
> *Subject:* R: Comparing NT Backup speeds
>
>  
>
> Yes the DAT device is local.
>
> Not applied registry tweaks
>
>  
>
> *GuidoElia*
>
> *HELPPC*
>
>  
>
>  
>
> ------------------------------------------------------------------------
>
> *Da:* Michael B. Smith [mailto:[EMAIL PROTECTED]
> *Inviato:* martedì 7 ottobre 2008 13.05
> *A:* NT System Admin Issues
> *Oggetto:* RE: Comparing NT Backup speeds
>
> Have you applied the "standard" registry tweaks to increase the 
> ntbackup buffer size?
>
>  
>
> Is the dat-72 locally attached?
>
>  
>
> Regards,
>
>  
>
> Michael B. Smith, MCITP:SA,EMA/MCSE/Exchange MVP
>
> My blog: http://TheEssentialExchange.com/blogs/michael
>
> Link with me at: http://www.linkedin.com/in/theessentialexchange
>
>  
>
> *From:* HELP_PC [mailto:[EMAIL PROTECTED]
> *Sent:* Tuesday, October 07, 2008 1:59 AM
> *To:* NT System Admin Issues
> *Subject:* Comparing NT Backup speeds
>
>  
>
>  
>
> In a network 10/100/1000 copper a NT backup of the complete server to 
> a Qnap device RAID-1 takes about 9 hours with verify (about 50GB) 
> versus DAT-72 with separate card that takes half the time .
>
> Should be considered normal ?
>
> TIA
>
>  
>
> *GuidoElia*
> *HELPPC*
>
>  
>
>  
>
>  
>
>  
>
>  
>
>  
>
>  
>
>  
>
>  
>
>  
>
>  
>
>  
>
>  
>
>  
>
>  
>
>  
>
>  


~ Finally, powerful endpoint security that ISN'T a resource hog! ~
~ <http://www.sunbeltsoftware.com/Business/VIPRE-Enterprise/>  ~

Reply via email to