Recent content by BackupProphet

You can maybe flash the bios to supermicro, without using CH341A share_id=06s_oTeWg_iIRx5phMh5r&utm_medium=android_app&utm_name=androidcss&utm_source=share&utm_term=1 Scroll down and you will see a 3 week old post.

There is a great deal here: Nutanix NXS2U2NL 12G600 Server NX-8235-G60-5120-CM 4x XEON GOLD 5120 2.2GHz 14co | eBay Chassis doesnt seem to use standard Supermicro Rails, so I cannot use them in my rack. And Im not interested in running these home.

Still available. I think these are fairly priced. They include rails, which is getting harder to find for a decent price.

Btw, this is common output for SAS hard drives. Which I actually prefer. Because monitoring of the Error counter log, will quickly tell you when a drive is bad. When the numbers there increases, its means your hard drive spend extra time on error correction which slows down everything. Also...

You have: Elements in grown defect list: 0 I think Elements in grown defect list is the same or similar as reallocated sector count.

It's not that special? You only need Cascade Lake or newer Intel Generations? For home lab, if youre stuck with a single Epyc system is understandable. But today, most can afford a cascade lake system which is just a few hundred bucks.

That is u.2, the other was HHHL. Either you already have u.2 bay available, or you need an adapter. I am suprised that these goes so fast, why are not more people looking into NVDIMMS? They have 100x lower latency compared to Optane SSD's. And are also cheaper per GB. Is it because you need...

Consumer hardware is really bad for many things, especially the SSD's are a joke. Samsung 990 Pro is among the slowest SSD's out there: source: Solidigm P44 Pro Linux Performance Review - Phoronix However, since that is a consumer SSD test, there is no guarantee that faster SSD's honor FLUSH...

From the model number, looks like HHHL

He also has 375GB available for about same price per GB. This is a good price that made me seriously considering buying a few for my Apache Pulsar cluster, but the truth is, I do not expect much better performance. And I cannot think of another use case I can currently benefit from them. So a...

I have never had any issues with mine.

Everything is as it should be. But I see you are using iperf3, just use iperf2. Iperf3 is SLOW compared to iperf2. I also recommend testing with other tools like netperf/netserver nuttcp or GitHub - Mellanox/sockperf: Network Benchmarking Utility

whats the output of ifconfig nic_name on pfsense?

iperf usually have a bottleneck using single cpu core. You can enable threads to test performance.

https://drive.google.com/file/d/1UVIveh50JGJ-Bfxu8voE1ovLh01CYhwP/view?usp=drive_link