Topic: SME5.6 HPT370A RAID Enabled Driver

[Kelvin]

Hi All,

Does anyone have a working Highpoint HPT370A Driver disk with RAID functions enabled for SME 5.6 ?

The included driver in SME 5.6 appears to have been compiled with the NO_RAID option which means it will only function as a pure IDE controller and any drives connected are visible as individual drives.

I followed a thread on these forums from about 2 - 3 months back which did not focus on the hpt37x2.o module required for the HPT370A. However, when attempting to compile the driver myself, I've also encountered the same problems the previous group did but did not find a resolution.

If anyone has a working solution to share, I would be most grateful.

TIA !

Kelvin

[Paul van Dun]

Hi Kelvin,

The drivers I've compiled for the HPT370A are made on a dual pentium II so give it a try and let me know otherwise I will try to compile a non smp version.

What I can't do is loading the driver at boot-time like 4.12 or 5.12 and type accept-dd and as far as I know there isn't someone on this forum that has a solution to enter a driver disk at boot-time.
So maybe there is someone out there that can put me in the right direction of making the driver to load at boot-time.

Regards,

Paul

[Kelvin]

Thanks Paul,

Your driver got as far as mine does only as well.

I've done some reading on the net and apparently the problem isn't a trivial one. Aside from the fact we cannot seem to get the driver disk format right, others who have come up with workarounds appear to recommend upgrading kernels (an upgraded kernel is one of the reasons we have problems in the first place !), or recompiling the current kernel and making changes mainly to do with removing the currently installed highpoint driver which does not have RAID functionality.

Neither prospect looks appealing at this stage !

Cheers,

Kelvin

[Kelvin]

Hi Paul,

I just got a hold of another brand of RAID controller to evaluate. I checked their provided driver disk and found that is has support for Redhat 7.3 to Redhat 9.0. Looking into the modules file, it even has support for the updated kernel 2.4.18-5 that SME 5.6 is using. So, I eagerly tested this new card.

Guess what ? Same problem with the "Insert the 1 Driver disk" problem everyone's been having. Firstly, Grrrrrr ! Then came, Hmmmmm ... can't be that even the hardware manufacturers can't get driver disks right. So, decided to test this.

Setup with Redhat 8.0 standard downloaded version - Works fine with the provided driver disk.

Then tested with Redhat 7.3 standard download version - again works fine with the provided disk.

I come to the conclusion that this might be a SME only problem. Could it be that Mitel's customised setup stuffs up the driver thingy ? I wonder ... I'll report to bugs and see.

Kelvin.

[Paul van Dun]

Hi Kelvin:

When I issue the command "insmod hpt37x2.o" I get this message:

Warning: loading hpt37x2.o will taint the kernel: no license

But the driver is active when I hook up a HDD to it, so can I also get rid of this message or is this just something that the opensource driver does ?

Regards,

Paul

[Kelvin]

G'Day Paul,

From reading other people's posts of woes with the HPT controller, I think it's because of the non-opensource library object file that gets compiled into the final .o file. However, that message should not stop the driver from working.

As I'm trying to install SME onto the array itself and not just use the array as a secondary storage, I would like to be able to provide SME with the right driver at install time and have it install directly onto the array, which we can't do because of the "1 Driver Disk" error which everyone is getting. Like I said, it's not just the highpoint driver that's affected. Another manufacturer's disk (for a different non highpoint RAID card) also does the same thing under SME but not under RH 7.3 or RH 8.0 even though the disk has support for kernel 2.4.18-5.

Kelvin

[Kelvin]

Futher to this, I just received my reply from Mitel about this issue that I submitted to bug report.

I've copied part of the reply below :-

>Thanks for the information. Unfortunately, we do not officially
>support the use of driver disks. Some people have had success with
>them in the past, but we don't test or support that feature of the
>installer.
>
>If the hardware is not detected by the installer,
>we are unable to install on that hardware.

In other words, we're stuffed as far as getting other hardware to work with SME.

Anyone know how to fix anaconda and build a custom SME iso ??

Kelvin