The Hipath 4000 V7 is not available on the network, only on the atlantic port

[vorossz]

The Hipath 4000 V7 is not available on the network, only on the atlantic port. If I skip all the routers and only connect one notebook to the server, I don’t see anything either. (assistant, comwin, yast, etc.) You can't ping. anyone's idea?

thx!

 

[Moriendi]

Connect to 192.0.2.7 on the atlantic port, login as root. Use ethtool and ifconfig to check the current status of the ethports (is port configured, is it up, is it down). Check the configuration XML in /var/opt/firstinstall to see how the ports should be configured and confirm it is correct. You can generate a new install XML from the current config by using the firstinst script in the same dir. Run "./firstinst-netw.sh -w" to create an install XML file which contains current config - then make sure the ports work like they should. If portal IP is down (no connect to 0.2.7) then try to the Assistant on 192.0.2.5 with engr login, then from there ssh to another internal IP of the 4K and do as above. You do not mention if duplex or simplex.

 

[vorossz]

Duplex. This happened a couple of years ago, but it didn't make sense for the customer to spend on repairs. Then I switched controllers manually. But now for some reason the error happened again. Now there are two controllers with the same error.

Thanks for the description. Can this be done during operation? The machine cannot be stopped during the day.

 

[Moriendi]

If you have two processors showing the same error it doesn't sound like the customer needs to spend on HW - it sounds like a configuration error.

Nothing I described is service affecting because you are only reading/checking config.

As well as Atlan working, Corosync then must also be working, for the duplex. Quick ping test from atlan would confirm but if it wasn't you would have heard about it by now, especially if the processors switched over OK.

Is there any IPDA? IPDA is configured to use same eth ports as the admin lan, or different? Is anything bonded/VLANs?

I would regenerate the firstinst file like I described above, and confirm it is correct.

If you do need to reconfigure the lan, the correct way is to reconfig the XML, and then use recover_4K to push the changes in, but that will stop the openais service (no telephony). The portal reconfig works for some things, but it would still cause an outage.