Sorry it took me a while to get back to my Rockstor project. In the past week I got another mother board that was known to work on Rockstor and tried to set it up as a Rockstor system. The exact same problem occurred with the new mother board. Given that this was a known good board, I turned my attention from the machine to the network. I tried to install the latest stable release on the new board after connecting it to an old Netgear 10/100 switch that was in turn connected to my Dell switch that is the heart of my network. The installation went flawlessly. I tried the TB250-BTCPRO board going through the old Netgear switch and it went flawlessly as well. I have never had any problems in the past with the Dell switch. It has some fancy features like POE, Qos, and virtual networks but none of that has ever gotten in my way before and I don’t think it has now. It has been totally comparable with everything I’ve hooked up to it until Rockstor. It has worked with PCs running centos, ubuntu, the latest OpenSuse, winXP, win7, and win10. It has also played nicely with the TB250-BTCPRO board with win10 and OpenSuse. It has never had an issue with various IP devices such as WiFi access points, IP cameras, and IP Phones. It just does not like Rockstor.
My take away from this is that Rockstor OS/Driver is working with the NIC hardware properly and the problem is that the Rockstor OS/Driver is not working with the protocols that the Dell switch wants to use.
Now that I can access the logs, I can send any logs that anyone may want to see. In the mean time I am looking into updating the firmware in the Dell switch.
Is there anything else that I should look at? I did look at the lspci -v command in the actual Rockstor installation. I’m not sure if it showed any problems or not. I’m guessing that at this point with the system working through the old switch it will not show any errors in the PCI bus but if you want to see the output I will try to figure out how to retrieve it from the main console.
Looking forward to hearing anyone’s ideas on what to do next. Maybe a setting in the switch or in the BIOS of these boards. Thanks for all your help so far.
OK, looks like no one had any ideas. I have been thinking about it for a while myself. Here is the experiments that I have run and the results.
First off, I called Dell and got some help installing the latest version of firmware onto the switch. This did not seem to do anything. I’m thinking about calling them next week to see if they can help me now that I have some more info on the problem.
Now for the experiments and results. What I have found is that I can install Rockstor v.15.5 if I place an old 10/100 MB switch between the install system and the Dell GB switch but Rockstor v.15.6 will NOT install in either configuration.
Once the Rockstor OS is installed and I switch to a static IP on the Rockstor System, I can plug the Rockstor system directly into the Dell GB switch and it works fine.
So, I have deduced that the problem is in the DHCP part of the install where the system is attempting to get an IP. My DHCP server is in the gateway server and is connected directly to the Dell GB Switch. Rockstor cannot get an IP if it is connected directly to the Dell switch but can if it goes through the 10/100 switch. So I’m thinking that there is a problem with the Rockstor system driver talking to the Dell GB switch but if it talks to the slow switch first, the message gets through. I have an instance of Kali Linux installed on a machine and I am willing to attempt to capture the packets in between the DHCP handshake if someone can walk me through it.
My other idea is to examine the driver in the Rockstor system for a problem. Again, can someone walk me through it.
when running 'lspci -v it shows you which kernel driver it’s using. I am only wondering that if a vanilla 15.6 install works, whether that particular kernel module is missing, using a universal driver and then runs into problems connecting. Or, for the setup that works installing Rockstor, which kernel module/driver is then used and try to figure out whether that’s missing from the base install to which it could be added.
But, not being a network expert, that might also be the wrong direction to go into.
I’m back… I’ve been busy talking to Dell and running packet captures on the communications between the Rockstor system and the switch. I’m pretty sure that they are going to exonerate their switch but I hope that in the process we’ll uncover some clues as to what is going on here. I’m going to get a bunch of packet captures from the experiments that I’ve been conducting together today or tomorrow and send to Dell. So hopefully by the end of the week I’ll get something back.
As for lspci -v, I installed a fresh copy of 15.5 and tried to run it. It gave me the “no such command” response when running it from the command prompt on the server machine. I have also gotten the same response for the commands ifconfig, and ifstatus eth0. I’m strongest primarily with bash on centos and then secondarily on Ubuntu. Not sure I’m even using the right commands for OpenSuse. What commands should I be using to retrieve the network info? Or is there a problem with the system? Also, it really would be nice to know set a static IP on the command line.
As for the 15.6 vs the 15.5, I determined that 15.6 was not working as well as 15.5 for some reason. At this point I have done so much ‘stuff’ with these machines I cannot remember what the reason was. I can go back and try to install 15.5 if anyone wants to know the issue.