Red Hat Hardware Certification Submitted Product: Red Hat Enterprise Linux Version: 4 Make: CPC5505 Model: B1 Vendor: Category: Server Reporter: jude.moersdorf Kernel Version: kernel-2.6.9-34.EL File Uploaded: rhr2-B1-Nahant_Update_3_results-1.noarch.rpm https://bugzilla.redhat.com/hwcert/data/rpms/feda9c6784e6aaad919f6ee9a2ba7ebb/rhr2-B1-Nahant_Update_3_results-1.noarch.rpm 0 PASSED USB PASSED INFO 1 PASSED INFO PASSED VIDEO PASSED STORAGE FAILED NETWORK PASSED CORE
Jude, Pretty good start; below are the remaining items I see from the product description provided on the updated url that we need to cover before we're squared away. Oh, and welcome to the program. -Rob The Network test failed; my guess would be that eth1 was not configured. The USB test was only configured for a single port; I'm assuming this is the USB 2.0 front panel port; the specs mention that 2x 1.0 ports are routed through cPCI Also routed through cPCI is an ATA33 controller (the ATA133 controller was already tested) Representative hardware is also required for all certified system; most likely a cPCI frame and blade would be required.
I'm not sure how the NETWORK test works, but both ethernet devices are usually configured for DHCP and they are on the same subnet, etc. Is this a problem for the tests? With many of our boards, signals can be routed to an optional card that can be plugged into the rear. How does this affect certification? If we certify just the CPU board, then we can't test some of the functionality that is routed to the rear. If we use a card in the rear, does this change what we are testing?
For the NETWORK test it is actually ok to have them both on the same subnet as the test brings them all down and then turns them on one at a time to perform the test. For the cPCI routing, it will depend on the specifics, my assumtion is the cPCI was simply trace routing where some set of wires equated the traces of the 2 USB ports and at some point that would be wired to a port. If that's the case then we should test for complete coverage, if on the other hand there's a concentrator or other hub/router/switch something(s) then perhaps we should discuss those in detail.
Created attachment 128366 [details] CPC5505 Manual
From what I see in the manual it appears the correct course of action would be to test the rear routed functionality of the CPU board but not any additional function that may be provided by the option modules (unless it is more desireable to include them in a sort of "CPU Board + XXX" certification).
Sounds good. I was out most of last week and will be getting on this stuff again this week. I believe that you were correct about the network configuration not being set up in the ifcfg-eth1 script for the second ethernet adapter. How long do you guys need the boards/system for? For this product, time might not be an issue. We are attempting to get a new product certified and we have very few production boards and 2 GB memory modules are very hard to come by...customers seem to take priority over engineering. Would it be possible for us to send a chassis with two different boards to be certified? One would be the CPC5505, the other would be our latest board the CPC5564. Thanks, Jude
> How long do you guys need the boards/system for? Representative hardware is required for the lifetime of the cert. > Would it be possible for us to send a chassis with two different boards to be certified? I presume this would be a chassis which takes both at the same time? If it's a single board chassis we may be able to work with that; however it limits our ability to provide support which is undesireable. Also to ensure expectations are set, we only require commitment/shipment of representative hardware before posting we won't hold the cert while the system is in transit as to not defeat the point of doing the testing before hand.
I'm going to go ahead and close this one out. We can re-open if needed.
Set it to be dumpest.