[THIN] Re: 2008 vs 2003 login times

  • From: "Anthony Green" <agreen@xxxxxxx>
  • To: <thin@xxxxxxxxxxxxx>
  • Date: Fri, 8 Jan 2010 11:37:05 -0000

Theres some group policy information in the eventlog but nothing like the
userenv debugging you got with 2k3.

 

Ive tried turning all group policies off which didn't really make a
difference.

 

I seemed to get better performance when disabling client mappings,
especially printer mappings, but nothing to shout about.

 

Am I gonna have to accept this and just say "Sorry, thats 2008 for ya, weve
upgraded you to the latest OS and given you a slower, but much more reliable
logon"

 

Thanks,

Ant.

 

From: thin-bounce@xxxxxxxxxxxxx [mailto:thin-bounce@xxxxxxxxxxxxx] On Behalf
Of Andrew Wood
Sent: 08 January 2010 11:20
To: thin@xxxxxxxxxxxxx
Subject: [THIN] Re: 2008 vs 2003 login times

 

One thing I noticed in deploying XA5 was how much extra resource it seemed
to use *before* any punters were on. Did you do some load testing before
your role out?

 

You could do a logon trace - maybe look at the timings ? if you've still got
a w2k3 box you could run a userenv.log trace w2k8 sticks the trace into the
event log iirc. Maybe there's a registry/script change/update that's hosing
the logon?

 

From: thin-bounce@xxxxxxxxxxxxx [mailto:thin-bounce@xxxxxxxxxxxxx] On Behalf
Of Anthony Green
Sent: 08 January 2010 02:25
To: thin@xxxxxxxxxxxxx
Subject: [THIN] 2008 vs 2003 login times

 

I've done a few upgrades recently moving people from 2003 to 2008 SP2 and
latest Xenapp.

 

Nearly every time the users have noticed a longer login time.

 

One company went from 7 secs to 14 secs (i know it's not a long login time
but it has doubled!).  We basically replicated their Windows 2003 setup
(xenapp\DCs\File\Print\SQL) but have everything on Windows 2008.

 

Most things work ok but the login times are killing me, I'm having to
explain why this is without really knowing why there is such a difference.
Anyone else notice this?  Any tips?

 

Thanks,

Ant.

 

Other related posts: