asterisk users - Sep 2004 - Asterisk server suddenly fronzen after many zapter errors

Hello all
We are running an Asterisk server using 24 extensions with an AB1 
channel bank and a T1 digium card, plus a TDM400 with 4 fxo modules 
(using only modules 2,3,4).

We have been using the pbx with this configuration for some time 
(asterisk 1.0rc1,1.0rc2).
A few weeks ago we have installed a mirrored disk (raid 1) in this 
server and we have been having some problems with sporadic noises, echo 
in the line and with high missdetection on dtmf tones.
Based in recomendations from this list an from digium support; last 
friday we made some changes in the hard disks settings (using the bios 
and hdparm) trying to fix it.

Now we have the problem it is the second time in this week we get 
thousands of warnings like these in the console:
Sep 30 12:15:38 WARNING[1228172208]: chan_zap.c:3766 __zt_exception: 
We're Zap/9-1, not 85"@85"@n**
Sep 30 12:15:38 WARNING[1228172208]: chan_zap.c:3766 __zt_exception: 
We're Zap/9-1, not 85"@85"@n**
Sep 30 12:15:38 WARNING[1228172208]: chan_zap.c:3766 __zt_exception: 
We're Zap/9-1, not 85"@85"@n**
Sep 30 12:15:38 WARNING[1228172208]: chan_zap.c:3766 __zt_exception:     
-- Hungup 'Zap/3-1'
Sep 30 12:17:54 WARNING[1228172208]: chan_zap.c:3766 __zt_exception: 
We're Zap/9-1, not Zap/2-1
Sep 30 12:17:54 WARNING[1228172208]: chan_zap.c:3766 __zt_exception: 
We're Zap/9-1, not Zap/2-1
Sep 30 12:17:54 WARNING[1228172208]: chan_zap.c:3766 __zt_exception: 
We're Zap/9-1, not Zap/2-1
Sep 30 12:17:54 WARNING[1228172208]: chan_zap.c:3766 __zt_exception: 
We're Zap/9-1, not Zap/2-1

and finally the server gets totallty frozen (not only asterisk) and the 
only way to recover is a full reboot.
Not sure how or if this is related to the new harddisk settings but we 
made the changes last friday and this (seriuos) problem happened tuesday 
and thursday (today) this week.

The settings we have now in our harddisks is using udma2 with the 
following settings as show with hdparm:
 dev/hda:
 multcount    = 16 (on)
 IO_support   =  3 (32-bit w/sync)
 unmaskirq    =  0 (off)
 using_dma    =  1 (on)
 keepsettings =  0 (off)
 readonly     =  0 (off)
 readahead    =  8 (on)
 geometry     = 4865/255/63, sectors = 78165360, start = 0

Does anybody have any suggestions or any idea in what could be the 
problem or how to fix it?
Do exist any experience in problems with digium hardware and asterisk 
when using linux software raid 1 in the server???

Many thanks in advance for any comment.
Luis