DL360 G1, SMH accessible only from localhost

[ETP_Vancouver]

I have a DL360 running Win2003 sp2. All software and firmware is up to date.

Locally, I can access the SMH via localhost or 127.0.0.1 (ie https://localhost:2381).
However if I use the server's IP address (ie https://192.168.1.108:2381) I get "The page cannot be displayed". If I didn't know better, it seemed as if certain ports were being blocked.
Another symptom is that the NIC connects only at 100Mbps and not at 1000Mbps.
Everything else seems to work fine.

The clincher:
When I transfer the disks to another box, everythig works fine! When I insert the disks from a working server, the SMH breaks.
I've replaced the memory, and reset the NVRAM.
This looks like a hardware fault on the system board. Trying an add-on NIC made no difference. HP diags comes up clean.

Have I missed anything?

[KarloChacon]

hi Petro

let me see when you transfer those HDDs to another server - same server model ??? -, and you use the IP from that server to access SMH you got the page working?

also what is your NIC card in that server DL360 G1 this one:

Two Compaq NC3163 Fast Ethernet NIC (embedded) PCI 10/100 WOL (Wake On LAN)

regards

[Blazhev_1]

Hi Petro,

are you able to succesfully ping the 192.168.1.108 from the host you try to reach SMH? If yes, then can be a browser security/certificate problem. Did you try from more than one server?

For the NIC speed problem check specs if it supports 1Gb, update firmware and driver and then from NCU check if the speed is set to auto- can be that your switch or remote host are using the same speed and are negotiating 100Mbit with this server.

Regards,
Pac

[Anift]

Try adding the Site address https://192.168.X.X:2381 as trusted site in the browser of the system from where are u accessing the SMH

[Mi6t0]

What about the SNMP configuration? How are the public/private rights configured?

[John Marinov]

If I understand correctly, you have taken the disks from one DL360, inserted them into another identical system, and the problem stayed with the hardware.

So on the problem hardware, the problem is between the server and the switch. One way to prove this is to use a crossover cable -use a laptop with a ethernet crossover cable and eliminate the switch completely.
From a workstation, telnet to your server on port 50000. Do you get a session, or does the connection fail (so are we getting a timeout because of TCP retries).

It is possible that the connection between the server and the switch is faulty, so check the ethernet statistics for faults.

Negotiation is sometimes the problem. Get your comms people to set the port to 1GB, full duplex, and do the same on the server.

Good luck.

[ETP_Vancouver]

Thank you all for you responses. I've read them all and will try to answer.

I removed the disks from the "bad" DL360 and installed them in an IDENTICAL "good" DL360.
Both servers' integrated NICs are HP NC7782 Gigabit Server Adapters.

I tried switching network ports & cables between the two servers. Made no difference. When forcing the "bad server" nic to 1Gb it refuses to connect to anything. The "good" server connects at 1Gb with either port.

The "bad" server is able to authenticate with the AD, respond to ping, and I can connect to it via remote desktop. From console, the SMH is accessible via localhost or 127.0.0.1. The SMH is *not* accessible when using the server's IP address - even from console. In this case, the browser connection should be looping back before it even gets to the NIC.

As soon as I move the disks to the "good" server, the SMH is accessible from console using the server's IP address, as well as from external client. I tried using the network port/cable from the "bad" server and everything still worked fine.

[Mi6t0]

Could you check if the revisions of the systemboards are different? Can you please note them here?

Is there some third-party hardware in the "bad" server?

[ETP_Vancouver]

I can't open the servers right now to check board revision levels. However, they are the same part #, and the SN's differ by only 4. They were probably very close on the production line.
Every SIM report and status shows both boxes being identical (ie firmware, microcode, drivers etc)

Neither server has any 3rd party hardware. Both are configured identically, with no additional cards.

I should add that when I cleared the NVRAM, the SN remained intact, even though the docs say it gets wiped. I chalked it up as an "undocumented feature".