A
andrewsplatt
We have an access 97 MDB that we had on an NT4.0 server. We have
about 30 wrksts that have mdb front ends loaded locally to access the
MDB backend. All was great until this evening when we decided to
ditch the 9 yr old NT 4 server and replace it with a Dell Dual Xeon
2003 R2 server. All data copied form the old server to the new server
was perfect. Performance increases were everywhere when we talk about
network thru put. However, once we try to get the workstations in to
the database on the new server, performance leaves and QUICK !. To
compare, the old server/workstation combo would take about 5 seconds
to open the database and about 7 seconds to perform a filter. Now on
the new blazing fast server we open the SAME database on the same
workstations in about 15-20 seconds and a filter takes over 2
minutes. Again, the workstations have not changed. The front end
still lives in the exact same place on the workstations and the
backend lives on the same mapped drive (just a different server). I
have read about a TCP ACK timeout that could be the cause. I have
also read about the ever present OpLock issue. Can someone who has
experienced this please help. The server needs to be up in 2 days so
converting to SQL is not going to fix the issue right now. Possibly
in a few months, but not immediately.
Thanks
Andrew Platt
about 30 wrksts that have mdb front ends loaded locally to access the
MDB backend. All was great until this evening when we decided to
ditch the 9 yr old NT 4 server and replace it with a Dell Dual Xeon
2003 R2 server. All data copied form the old server to the new server
was perfect. Performance increases were everywhere when we talk about
network thru put. However, once we try to get the workstations in to
the database on the new server, performance leaves and QUICK !. To
compare, the old server/workstation combo would take about 5 seconds
to open the database and about 7 seconds to perform a filter. Now on
the new blazing fast server we open the SAME database on the same
workstations in about 15-20 seconds and a filter takes over 2
minutes. Again, the workstations have not changed. The front end
still lives in the exact same place on the workstations and the
backend lives on the same mapped drive (just a different server). I
have read about a TCP ACK timeout that could be the cause. I have
also read about the ever present OpLock issue. Can someone who has
experienced this please help. The server needs to be up in 2 days so
converting to SQL is not going to fix the issue right now. Possibly
in a few months, but not immediately.
Thanks
Andrew Platt