Proliant Support Paq 6.30 for RH8

[Tom Mucha_1]

Hey all,

I could use some help, I'm stuck!

Before I start, I have updated the kernel to 2.4.18-27. All my packages are up to date too. The server is a Proliant ML370G2 2.8Ghz.

This is a new server, and I was going to try to install the "Proliant Support Paq v6.30 for Redhat8". I guess it does Compaq/HP drivers and all the management agents at once.

Well after I run that support paq, I get these printed out constantly to the console and /var/log/messages:
May 21 16:04:49 fs kernel: (scsi0:A:6): 40.000MB/s transfers (20.000MHz, offset 7, 16bit)
May 21 16:04:49 fs kernel: (scsi0:A:6): 40.000MB/s transfers (20.000MHz, offset 15, 16bit)

This support paq isn't really necessary, but I'd like to figure this out if possible.

TIA!

[Stuart Browne]

Did this only appear after you put the care-pack on?

These appaert o be kernel messages, of which usually only show up during the loading of file-system modules (i.e. only during boot-up).

I doubt that these modules are being loaded repeatedly. I also don't beleive the care-pack's make any calls which would cause the kernel to throw these up.

I must admit I've not tried the latest kernels under RH8 (they break ABI which we need here), so I've not seen similar behaviour.

[Tom Mucha_1]

They would show up once upon loading before the Support-Paq (during boot). After the support paq, the show up constantly.

I was doing some digging to see if the support paq only supported a specific kernel, but haven't found anything MUCH yet (looking from home-- scarry!).

Tommorow when I get some time I'll try to undo the damage some how!

[Michael Garner_1 · Accepted solution]

Some components in the PSP for RH80 are kernel dependent while others are not. To determine what kernels are supported by each component, you can look in the .xml file associated with each RPM in the PSP. In the section of the XML files, you will see a listing for each distribution that a component supports. Within a given distribution, you will see a section called . Under this section, will be a number of SubType="" entries followed by values. A SubType="6" entry is a minimum kernel supported entry. A SubType="7" is a maximum errata kernel supported entry. These values are set based on developer input for the given components. Changing them will not necessarily cause the RPMs to install because the individual RPMs may not support later kernels. A value of "2.4.99-99" means that any 2.4 based kernel is supported.

None of the storage drivers in the PSP 6.30 supported the 2.4.18-27 errata kernel, so none of them would have installed.

You are also correct that the PSP deployment utilties don't issue any kernel calls. However, several of the components when they install, ie hpasm, e100, e1000, and bcm5700 do rebuild themselves and insert into the kernel. However, none of these would affect the fs kernel.

From the information you provided, it appears that one of the drives is not working at a slower speed that the others on the SCSI bus. This could be caused by quite a few things, including a tape drive on the same bus, bad cable or terminator on the SCSI bus, a drive going bad, a drive or adapter needing a firmware upgrade, or any number of things.

Does the problem go away if you reboot the server?

What is the configuration on your SCSI bus?

If you connect into the Storage Agent through the web browser (use https://127.0.0.1:2381 to connect to the web server embedded in the agent software), what does it report?

You can also use the new Insight Manager on-line Diagnostics for Linux (hpdiags) located on the SS CD (/compaq/csp/linux/hpdiags-3.0.0-75.i386.rpm) or on the web to run diagnostics on your storage subsystem to determine what the cause of the problem is as well. After installing the hpdiags RPM, you can use the embedded web browser mentioned above to run the diags on the local server.

Hope all this helps.

Michael Garner
Linux PSP Developer

[Tom Mucha_1]

Sorry for never replying, I never got a notification anyone else posted.

But anyways, I found out the message is coming from the the Storage Agents (just did another install). If I use the 6.20 agents, everything is fine. So once I stop cmastor, everything work fine.

I'm going to try that diag (searching for libGLU.so). I'll let you know what I find out.

Oh, and I tried this on another identical server and got the same results (ML370 G3 2.4 with an internal 20/40 DAT on the embedded SCSI controller).

Thanks!