Today, out of nothing akl ports died, including the SFP+ ports, The strange thing, the switch looks as it is booting.

Now I wonder, is there a console port on it? Remind all ethernet ports are DEATH.

After power off the switch does start and do its usual boot:
- blue led goes kn
- fan spins up to.msximum
- fan starts steppjng from slow to fast
- blue led goes off
- fan goes to normal speed
- blue led goes on again
But at this moment nothing haplens, interfaces stay down.

Last change on switch, approx october when i added my new NAS on port 7&8 in a trunk.

I want to diagnose its console boot, because i cannot believe a device dies that easy.

Serve The Homes does nice job showing the insides:
https://www.servethehome.com/mikrotik-c ... -switch/2/

1. You can try and recover the Device by loading the backup RouterBOOT
Mikrotiks Website (https://help.mikrotik.com/docs/spaces/U ... 0-8G+2S+IN)

2. Alternatively, if the Device Itself isn`t defektive sinply reinstall the System with NETinstall.

Only ethernet ports? Can you try SFP?

And the usual:
* Restore factory config.
* Do netinstall.
In case they're disabled or installation is somehow corrupted.

The reset button has these functions:

Tried:
Hold this button during boot time until the user LED light starts flashing, and release the button to reset the RouterOS configuration.

Did not do it's thing. All ports death eth1 - eth8 and sfp+ 1 and sfp+2

Tried:
Keep holding the button for 5 more seconds or until the user LED turns off, then release it to make the RouterBOARD look for Netinstall servers. The first Ethernet port is used for the Netinstall process.

.
Netinstall did not work. Port 1 stays dead.

Will try:
Regardless of the above option used, the system will load the backup RouterBOOT loader if the button is pressed before power is applied to the device. Release the button before LED begins to flash, to only load the backup RouterBOOT without resetting. This is useful for RouterBOOT debugging and recovery.

Will let know.

My seller wants me to send device back to warehouse. And does not acknowledge until they receive it, it is defective. I do not know if this is a standard practice, but this makes me worried about waranty at all and use Mikrotik in a business environment.

Ok, i started over and did use different setup with a small switch between my USB network adapter and Mikrotik Switch Port 1. And surprisingly, it did reboot and was seen with netinstall. So the switch is functioning.

I don't understand why all ports are dead, lucky had a configuration backup and all ports worked fine with new firmware. Now it is in stability test to see what goes on.
The strange thing, it died exact at 12:00 PM.

Now I still have to figure out what happened and why the switch died. I disabled the automatic firmware update so that is not going to be an issue.

There are many possible reasons for device to misbehave. Unfortunately one of them is (invisible) configuration corruption which is also saved in binary backup. If such backup is restored on (newly installed) device, corrupt setup is back in place and waiting to screw things. So if the problem will show again, the best way round it is to create textual export of config (command /export show-sensitive file=anynameyouwish), fetch file off device, netinstall it and then apply config via CLI (or GUI, translating config items to many mouse clicks). Textual export doesn't contain user data, certificates and SSH keys (possibly few more things), you'll have to re-create those things in different ways.

.... /export show-sensitive terse file=anynameyouwish),,,

That helps to paste config line by line.

I always make an export backup, because I like it to have a readable version of it.
I started to donor that way because I had issues with ASUS where restore settings after upgrade really bricked stuff. Also from a professional point of view it is handy when you want to document things why you do things you did.
Now my home setup is to small to put things in Git to make version history so I stick with just text based config backup.
So far so good it is still running. But I still wonder what went wrong. And sadly nothing logged of interest on my SYSLOG server.

Sometimes people plug in power, check led to confirm working device and disconnect it. If you have extremely unlucky timing and unplug it during the first boot, when it applies default settings and writes this configuration to disc it could corrupt configuration. But it happening exactly at 12:00 on its own, sounds like something time dependent was involved. Maybe a script or power switch-over?

I'd also check license with /system license print
With corrupted installation this can be corrupted as well and demo times out after 24h. If RouterOS time was not set timeout would happen at 1200.

I checked the license and the config. All looks great but I think some settings I had where not well for the device, although it might be a firmware bug or so. For now the device is stable.
My device is behind an UPS (which works well) and my home is behind an off-grid hub (with battery and solar) so no power interruptions happening
But I have to admit that I do not always did a shutdown of the device gracefully because it is never needed with the Cisco and Aruba stuff as long as you do a write config.

I checked the license and the config. All looks great but I think some settings I had where not well for the device, although it might be a firmware bug or so. For now the device is stable.
My device is behind an UPS (which works well) and my home is behind an off-grid hub (with battery and solar) so no power interruptions happening
But I have to admit that I do not always did a shutdown of the device gracefully because it is never needed with the Cisco and Aruba stuff as long as you do a write config.