| Title: | *OLD* ALL-IN-1 (tm) Support Conference |
| Notice: | Closed - See Note 4331.l to move to IOSG::ALL-IN-1 |
| Moderator: | IOSG::PYE |
| Created: | Thu Jan 30 1992 |
| Last Modified: | Tue Jan 23 1996 |
| Last Successful Update: | Fri Jun 06 1997 |
| Number of topics: | 4343 |
| Total number of notes: | 18308 |
Hi,
Quick question..
Customer is using ALL-IN-1 v3.0, and is trying to improve performance for
the File Cabinet Server. Has anyone tried setting global buffering on
partition.dat for performance improvement, and does it seem to make a
difference?
What recommendation should I give to the customer on the value of the
buffer size? In the manual it states that most .dat files should be set to
number of user+1, but partition.dat is not mentioned.
Thanks,
Julie
| T.R | Title | User | Personal Name | Date | Lines |
|---|---|---|---|---|---|
| 3825.1 | Is the FCS tuned | CHRLIE::HUSTON | Fri Jan 28 1994 13:25 | 7 | |
Is the customer using the default server configuration and parameters?
These are set for a rather small system. If they are bigger than
50 users (I think) then it needs tuning by adjusting cache sizes etc
--Bob
| |||||
| 3825.2 | KERNEL::OTHENJ | Fri Jan 28 1994 14:07 | 11 | ||
Thanks for the reply Bob,
They are hardly using the FCS at all, so do not have more than 30 users
connecting on the the FCS at the moment, so the default settings should
be O.K. They have global buffering set on all other .dat files though,
so would like to set it on partition.dat as well. Any recommendations??
Thanks,
Julie
| |||||
| 3825.3 | Nunber of index buffers | IOSG::DAVIS | Mark Davis | Fri Jan 28 1994 17:15 | 22 |
There may be a small benefit in putting some global buffers on
partition.dat - because users will open this file when they enter
ALL-IN-1. You can check this with $SHOW DEVICE/FILES/NOSYS OA$DATA and
you will see many users accessing it.
However rather than put the number of users + 1, a smaller number would
be better - the number of index buckets + 1.
The number of index buckets can be calculated by
$anal/rms/stat oa$data:partition.dat
Divide the Count of Index Blocks by the Index Bucket Size to get the
number of Index Buckets.
Any performance improvement is unlikely to be large.
Mark
| |||||