@benjojo Perfectly backpack sized ​​

@benjojo "acquired"..... so who lost connectivity? ;)

But good to see that I am not the only one transporting network/servers in a backpack.... or did you know how versatile and waterproof those blue IKEA bags are? ;)

@benjojo this is a nice line of gear. which OS do you plan to run on it?

@benjojo looks so spacious and well designed compared to the (Trident3) Dell S5212F. Additional boards galore, twice the fans due to the silly fan placement, etc.

@benjojo this looks like a half-width switch to me. Is it one of the Mellanox ones you posted in a photo some time ago?

If so, it could be that this is for an option to "stack" two of these units next to each other.

@benjojo I would guess its just for testing the board after it rolled of the factory. Or at least test the asic it self. Given it seems the board is used for both the SN2100 and SN2010 but would need to pull one open at work to check

@erikk @benjojo well what happened when you plugged a QSFP in? 😆

@benjojo random test pins wouldn't be able to carry high-speed signaling… and there is at least one high-speed thing you might want access to for factory programming: PCIe

@equinox @benjojo Yes altho qfsp has enough lanes to carry at least PCIe x4

@erikk @benjojo yeah, could be anything between ×1 and ×4 if it is PCIe

(funnily enough, for a "full" PCIe ×1, you'd still need a QSFP connector instead of SFP; there aren't enough pins on SFP to carry the PCIe clock and auxiliary signals. It can be made to work without those, but depending on details that can be annoying…)

@benjojo also, although the position is a relatively strong argument against it, this could be breakout from the SoM

@benjojo I sort of understand why so many things panic-ramp the fans to 100% when they start powering up, but boy do I wish they didn't do it. Especially the things with BMCs. Dear BMC, you already have a ton of sensors, can't you tell that the system is not going to overheat just because you started the CPU?

@cks @benjojo I want to agree.

But there are weird failure modes where the CPU might be hot. Things like crasooped from thermal problems and not safe to start slow.

But I feel like analog electronics sensing temperature and ramping speed up will work independent of the firmware.

@cks @benjojo There's unfortunately often a delay between when the BMC powers on the host processor and when the interfaces by which the BMC can read the temperature of that processor (e.g. PECI on Intel platforms or SB-TSI for AMD) actually come fully online. The fans are usually on the same 12V power rail as the host and hence turn on when it does, and lacking a valid temperature reading from the host CPU, going into failsafe mode is the...well, safe option. Logic like "if we just turned it on right now on it's probably not very hot" runs into problems if it had recently been on and is still holding a lot of residual heat...you could potentially get into tracking more history to disambiguate that in turn, but then you're suddenly a lot more stateful than you were which gets messy and fragile (especially considering that the BMC and the host can both reboot independently of each other), and it's ultimately just a lot simpler and less error-prone to make it (relatively) stateless and err on the side of not cooking things. And of course since most servers end up situated in places where there usually aren't people around to hear them, acoustic noise optimization is typically pretty low on the list of priorities.

(Yes, I work on BMC firmware.)

@zev @cks @benjojo I have plenty of beefs with BMCs, especially the recent trend toward flashy colours and automation in web interfaces and the rarity of CLIs other than IPMI, but fan speed ain't one of them.