Search results

For what? NTP via public peers is good enough to keep time synced within *very* small deviations. Sync all your local machines against a pair of hosts that are synced against upstreams and you're done...

enterprise SSD/NVMe without cooling is bound to end in (very) early failures just like it always was with most (especially high-rpm) server HDDs. if you need something for passive cooling, go for low-power desktop-/laptop drives, e.g. WD blue will run at very low temperatures. of course you...

we run micron 7400 and 7450s everywhere and they run fine. "loosing a partition table" also doesn't sound like a typical hard drive error... was this by any chance with some (pseudo) hw-raid that does some proprietary on-disk-format?

what exactly do you need if you say "L3"? If it's only some routing protocols and maybe PBR, that can be handled by almost all so-called "L2+" switches nowadays... E.g. the FS S3200-8MG4S-U would fit your needs with 8x2.5 + 4xSFP+ ports: https://www.fs.com/products/157457.html If you need/want...

you can just download the latest firmware for the 3650 series from Cisco Software Central, you just have to create a (free) account.

just get a standard quad carrier board for ~30$ and use bifurcation. no need to buy such weird crap...

Because a lot of desktop boards don't support PCIe bifurcation, which is needed to run e.g. 4 M.2 NVMe drives in one x16 slot "the right way". Don't know what they messed up in that product description (or in the actual product?), but if the platform properly supports bifurcation, there is no...

what you need for that use case is more IOPS - so spinning disks are already sub-optimal (read: bad), and RAIDZ doesn't give any improvements. Add a horrible protocol like SMB to the mix and real-life performance goes down the drain... ZFS *needs* to spread the load over multiple vdevs - so...

Have a look at the current MFC models from Brother. We have ~5 of the monochrome MFCs and also ~20 Brother printers and those things just live forever at minimal maintenance and operational cost. At home I still have a HL-2030 from ~2005 that works like a charm. Replacement toners are very...

just add the coupon code "WINTERPOWER" and you'll get another 10% off:

After an initial sync, how big are the deltas over one/several hour(s) or a single day? Syncing multiple times per day via that 38Mbps link should be no problem if you aren't writing multiple GB of new data every hour...

According to a bunch of search results, this error points to MSI being deactivated... Can you give me the output of 'sysctl -a | grep msi'? MSI(-X) is not in some legacy-mode or disabled in BIOS? And just to be clear: this is all bare metal? Not on some hypervisor and SR-IOV magic?

what's the output of 'sysctl sys | grep mlx' and 'sysctl dev | grep mlx'? Also I *strongly* suggest not using OPNsense or PFsense but vanilla FreeBSD. Especially the latter ones are using the development branch, which was NEVER intended for production use (and *will* break from time to time...

try setting the "sys.device.mlx4_core0.mlx4_port1" sysctl to "eth". I just checked some FreeBSD hosts with ConnectX3 NICs and I've also set that sysctl on those.

Usually Mellanox ConnectX support ethernet and infiniband, so you'd need to load the appropriate driver for it to be switched in the corresponding mode. The 'core' driver on its own doesn't register any device in the OS. No idea how Linux handles this, on BSDs to get it running in ethernet...

"molex to sata, loose all your data" :rolleyes: I'd strongly recommend using a proper backplane instead of fiddling around with multiple connectors per drive. this is bound to cause problems at some time and a master recipe for data loss.

console usually defaults to 9600 baud. wrong settings should at least produce some garbage - so if there's absolutely nothing, the board is very likely dead. if it is only a broken firmware, recovery via a USB drive might revive it. I'd also contact supermicro support, they are usually very...

if there's not even serial console output that thing is most likely dead... after a cmos-reset you should be able to see the POST output (and the ipmi address!) and entering the BIOS via serial console even if the VGA output is dead (or disabled via jumper - you checked that?). You should also...

fs.com For transceivers and cabling they are by far the best bang for the buck. We're using them everywhere and never had any compatibility issues. Very fast shipping, very fast and helpful support and in the very rare case of a defect (had 1 DOA transceiver out of around 150 we've bought so...

RaidZ loses at least ~15% (more realistic is ~20%, worst case easily up to 30%, especially with very few disks like in your case) to padding and meta/parity-data. So forget about those "usable space" estimations for RaidZ. As @mattventura said: if you have a special use case where you *need*...