| 1 |
Richard Watson <waty <at> bigpond.net.au> writes: |
| 2 |
|
| 3 |
|
| 4 |
> Hi - I'm getting vastly different performance results on hda (master) as |
| 5 |
> opposed to hdb (slave). |
| 6 |
> Any suggestions would be appreciated. |
| 7 |
|
| 8 |
I'd make sure you read the man page on hdparm. Certain actions can kill hardware, |
| 9 |
permanently. |
| 10 |
|
| 11 |
That said, what I do is set the hdparms in /etc/conf.d/hdparm. I believe you |
| 12 |
can test each drive, note the best (safe) parameters to adjust. Test the |
| 13 |
adjustments, and then make an entry, per hard drive in the |
| 14 |
/etc/conf.d/hdparm. Here's what works for me on one system: |
| 15 |
|
| 16 |
all_args="-d1" |
| 17 |
hda_args="-d1 -u1 -c1 -a256" |
| 18 |
|
| 19 |
|
| 20 |
Remember, any device that is seen as a /dev/hd* will be affected by |
| 21 |
settings like [all_args="-d1"]...... |
| 22 |
|
| 23 |
If get lesser performance when (2) different drives are on the same |
| 24 |
(ide/ata/eide) controller, then try a second pci cotroller card. |
| 25 |
or if both are on the same cable, and the MB or controller has a |
| 26 |
second port, add a second cable and move one drive to the other |
| 27 |
port/cable. This sort of performance degradation should effect both drives. |
| 28 |
|
| 29 |
HTH |
| 30 |
|
| 31 |
James |
| 32 |
|
| 33 |
-- |
| 34 |
gentoo-user@g.o mailing list |
Archives