🚀 go-pugleaf

Henrik Flindt <gpv_int@pip.dknet.dk> writes:


>In article <42l1dh$494@clarknet.clark.net>, <tsr@clark.net> writes:
> 
>> At one of our client's sites, we're ready to deploy a new (upgrade) of
>> probably either 4.1 or 3.12.  Since the existing disk subsystem is
>> duplexed (a pair of Adaptec AHA-1542's and a pair of Seagate 1.02 GB SCSI
>> drives), it is only logical that they want to go with a duplexed 
>> configuration again, rather than a RAID array due to cost considerations.
>> 
>> The equipment that was specified were:
>> 
>> 2 Adaptec AHA-2940 PCI host adapters
>> 2 4.x GB SCSI drives, either Seagate's HAWK (ST15230N) or Quantum
>> Intel Advanced/ZP ("Zappa") 75 MHz
>> 32-48 MB of RAM min.
>> 
>> Anybody here have any horror (or success) stories of using the AHA-2940's 
>> in a duplex environment with any of the above drives on a PCI bus?

>OH - YES come to me, and I will tell you horror-stories.
>I have experiencet quite a lot of problem with this combination ( I used some 
>bigger disk - but that shouldn't matter)

>The only thing you should be aware of is :
>- Make sure that you FS has a lot of memory 
>- Make sure, that you will get the newest firmware release in your Adaptec 
>controlleres. (AHA controllers have a (major) bug in them. When it comes to 
>intensive diskload on the server, it can cause the server to abend or the 
>disk's to deactivate).

>Good Lock !
>(My servers works fine now ! )

>Regards.

>Henrik Flindt

Interesting. I have a configuration with two 2940 too, both connected to
1 GByte IBM disks. Dont know the exact type.
Works well in principal but in contradiction to my old 3.11 server
each unnormal shutdown abends mirroring of the disks. Remirroring
is in my case a very slow process. For these comparative small disks
I need about 40 minutes for remirroring. In this time access to the
server is very slow. I have to mention the server itself: Pentium 100,
128 GByte of main memory, ASUS mainboard.
Is this slow remirroring a normal process? Is this maybe due to
compression?

Rainer Haessner
TU Muenchen
rh@linda.org.chemie.tu-muenchen.de