Condom reconditioning is the new fastest growth market.

@JaredBusch said in Who do you use for a 1U RackmountPC?:

@brianlittlejohn said in Who do you use for a 1U RackmountPC?:

Dell has 1U precision workstations. I'd look at those.

I thought they stopped making those.

They have a 3930 on the product page that is 1U and can be configured with 9th gen Intel CPUs.

Dell has 1U precision workstations. I'd look at those.

You have to use software RAID since NVMe is attached directly to pcie bus.

@scottalanmiller said in Getting up and running with ER-X?:

@pmoncho a friend of one of our customers got hoodwinked by the Crosstalk guys! They sold him some piece of junk that was literally less than a free phone system and billed him $5,000 for the privilege of getting screwed.

From what I remember watching some of his videos he likes to use the PBXact appliances.

@WrCombs said in Collision Domain - In POS:

Just replaced the first switch in the bar a week ago and it didn't help with the issue.

I've seen gigabit switches fail to negotiate speed correctly and had a 10M half duplex link.

Change the port the line from the back to the bar is plugged into on the switches to test out a bad port.

You could also run an iperf test from one of the up front pcs to the server and see what kind of bandwidth you are actually getting to the server.

I would put money on the actual line being damaged somewhere (bad crimp, punchdown)

One port isn't an issue unless you are saturating that one line, also it is all going down to one line when it goes to the server. If you saturate more than the switch can store and forward it drops the frame. It would be odd if that was the case giving the small number of devices running on the network. I would check and make sure the switches connected to each other are both negotiating at the full speed of the switch e.g. gigabit if it is a gigabit switch.

Does the tablet freeze up as well when the other machines do? If so, i would look at the switch in the back as the issue or the line going to the server.

If it is just the front machines, it may be an issue with the line itself, or the first switch in the bar.

On switches you can't have a collision, they are layer 2 devices with queues. Collision domains were a problem back in the day of hubs because 2 devices could send at the same time. Hubs operated on layer 1 and you would split collision domains with a device called a bridge, it would take packets and store and forward them between the collision domains. A switch is just a multiport bridge, each connection on a switch is its own collision domain since the device will store and forward for each segment.

Odds are your issue are a cabling issue, or a bad NIC on the server, or a bad switch/port on switch somewhere.

It also could be a software issue on the server causing the issues.

FreeOTP Authenticator from Redhat is another option.

I believe the ERX will offload some of the processing.