Thank you patient0,

I am humbled, and very grateful to have been taken to school today.

Thanks a bunch!

thumper23

@SteveITS Thanks - each port "being its own network" is a helpful bit of knowledge. I have a lot of learning to do so I'll move on from here, probably start a new post with more questions at some point.
Appreciate your insight and time!

Hmm, that is odd. Momentary press? 1s press? 3s press? All of them?

@andrew_cb said in Another Netgate with storage failure, 6 in total so far:

@SteveITS said in Another Netgate with storage failure, 6 in total so far:

The amount of disk space used by pfSense is typically relatively small so size isn't really a factor unless downloading large lists or data like the UT1 list which is over 1 GB to extract, when it updates. A larger SSD though would have more writing capacity, I'd expect, due to more unused sectors.

I think this is a critical part of the equation. Settings and packages that log more than the expected baseline, combined with the behavior of ZFS, lead to more storage writes, which go to a small storage device with TRIM disabled/not supported,

TRIM (or an equivalent such as DISCARD) are required by JEDEC standards as far back as 2010.

and a limited number of spare blocks. This quickly exhausts the approximately 3000 write cycles of the eMMC storage. By my calculations, 5TB is the approximate limit of data that can be written before an 8GB eMMC dies.

Yet you do not show these calculations, such that we may assess their accuracy. Specifically, I'm curious what you used for a Write Amplification Factor and how you determined same.

I actually came back to this thread to see if you had responded, but apparently not.

No me either but I'd quite like to find out! Are you able to make the failing config available to us to check?

@stephenw10 I've got a gazillion used m.2 sata drives, I'm sure one of them will work. :) Thanks!

It's a pretty standard socket. 2.5x5.5 barrel plug I believe. The only unusual thing is that it's locking.

@stephenw10 just upgraded my 6100 MAX to 24.11 and this has also appeared in my SMART widget. Have set to ignored but dont remember this ever appearing before the upgrade.

Opened an RMA, shipped it this afternoon, the eMMC presumably did fail. When I get it back I'm definitely moving off the internal drive and onto an NVMe. Thanks Netgate for the quick response.

I think it's pretty clear the eMMCs are causing repeated problems, and I'm sure there's a threshold where you should decide to do away with the "base" model and just automatically drop an NVME in these by default as you sell them. Doesn't have to be huge or expensive.

@Gertjan Sorry I wasn't super clear in my sentence but I was simply mentioning that no IP address was given and I cannot access the console so I was a bit dead in the water. I was able to use a third Micro-B cable which worked and I can see that the 1100 is not booting fully. I am in midst of trying to reinstall pfSense on the device but the first attempt via a flashed USB using etcher left me with a sub child process failed during the installation of packages. I'm trying a different USB thumb drive and will report back.

@stephenw10 Yes that seemed to be the problem; I tried a third Micro-B cable and it worked. Sort of silly!

@coxhaus I would love to given the positive experience I have had with the CISCO SG200 switches.

I would like to upgrade to at least 2.5G and everything I can find in CISCO products are overkill for my needs, even staying at L2.

@davebonnell said in Netgate 6100 replacement 128GB NVMe drive:

And the compatible drive should be 2280 size NVME and have a b+m key.

I believe the "2280 size" was intended to mean that it cannot be larger than a 2280. It doesn't actually need to be 80mm long.

I don't believe there is a mount for a 2230, but there is definitely one for a 2242 which is a common size and avoids the need for an adapter. If memory serves, there are mounts for 42, 60 and 80. YMMV.

@Nellies yeah tac is normally johhny on the spot.. I have gotten link for download in the past in couple of minutes, think longest was like 20 minutes and that was right when a version dropped and you know they had to have been swamped with tickets.

So your 3100 is back in business?

Hmm, well I guess I would test a client that isn't the switch just in case it has some quirk that prevents it using a VLAN correctly for management. But if you used that same setup on a 6100 it should work here too.

Otherwise a pcap at the client will at least show if it reaches it.

@CatSpecial202 I never got it to boot but basically Netgate support just said to send it back after I reported the blank console and the power button behaviour.

Did you have a 4200 or a 4200 MAX?

There isn't at the moment

The 4100/6100 kit will not fit. The 4200 doesn't have the same mounting holes.