flyride

The U-NAS designs are focused on being compact versus convenient disassembly. You will find you need to remove more parts of the chassis to do basic reconfiguration and/or maintenance than you would expect. You will also probably need to look at cable configurations for SATA and power. The 4-bay is easier than the 8-bay. I had to make custom cables and careful routing to make the 8-bay installation sanitary, although I admit I was setting up a fully solid-state power supply.

Don't ever subject your data to a migration event without a backup. These are probably Bad Ideas. If you installed DS3615xs or DS3617xs, you can turn on SHR, see the FAQ post. Normally I would say migrate the drives as part of the install but from DSM as old as you have, I really suggest you invest in at least one high capacity drive so that you can migrate to a new system while your old system is still intact. Then you can re-establish the RAID configuration you wish to use.

https://xpenology.com/forum/topic/28183-running-623-on-esxi-synoboot-is-broken-fix-available/

Easiest way for that is to add a hostname override for the nas on each 10Gbe client to point to the network servicing the port they are attached to.

If you are receiving a corrupt file error message, read this: https://xpenology.com/forum/topic/28183-running-623-on-esxi-synoboot-is-broken-fix-available/

How many 10Gbe clients do you have? I just use a dual-port Mellanox in the DSM server and DAC to the two clients that I want 10Gbe from. No switch.

You have some homework to do. I will point out two problems that need fixing: 169.254.x.x is the default IP given to a device that fails to retrieve an address via DHCP. So you do not have a valid DHCP server on your network, or the NIC cannot see the DHCP server to retrieve an address. It basically indicates a lack of communication. You can also manually assign addresses via the UI (static IP addressing). Modifying synoinfo.conf requires root access. In Linux and therefore DSM, this involves logging in (to SSH) and then elevating your access with the command sudo -i

My ConnectX-3 does not tell me any SFP/DAC information.

I have had the best results ordering direct with international shipping, but they do go out of stock on a regular frequency.

ARM won't work for XPEnology, no loader support for non-Intel. But plenty of people have done low-cost ITX based hotswappable NAS chassis. I personally use the U-NAS 410 and 810, but many choices on the market.

The local port is what your browser will connect to. Are you using bridge mode for the container?

Is there a reason not to have your local port match the container port? Only reason to change them is if you already have a service on those ports, which is unlikely.

Your motherboard has some limitations between NVMe and SATA SSD's and some overlap with SATA addressing. Not sure if any of this applies to your situation.

If you don't mind, please launch ssh and run this and post the results: lspci | grep "Class 0200"

Best results are with native SATA connections. USB drives can be hacked in, but it is fragile. You cannot easily use M.2 NVMe for "system" - DSM naturally installs to all SATA disk devices and M.2 NVMe is reserved for optional SSD cache unless virtualize your SSD or go through some significant further hacks. Remember the real value of DSM is the value-added features over basic file sharing - redundancy, self healing, replication etc which require proper access to SATA hardware. If all you need is file sharing, your proposal is the hard way to do it. The most effective XPe hardware solutions are platforms that look a lot like the native Synology hardware.

I haven't ever configured a system this way, but see no reason that you cannot reorder the drives to your heart's content.

I can't advise you on this. I use Jun's loaders and I'm quite security conscious. Thousands of other people have trusted Jun's loaders, but we don't know who it is, and don't have source code. I guess I am really just pointing out inconsistent security logic when your base environment is a deliberate hack. Would you trust a package from an entity you don't know, if it were signed? What's the difference? Anytime you are installing a SPK, you are giving root access to the application. When I have a choice, I prefer Docker, where at least I have some modicum of control over the operating environment of the application.

I'm not sure there is anything to do about it. The nvme binary is standard Linux accessing the kernel driver and is unaffected by the patch. I would guess that you would see the same behavior with a real DS918+.

Not sure what this means; DSM supports hot-swap.

How do you know the loader is safe?

So you have three storage pools?

Frankly, these are things you ought to know before you deliberately swap out drives as you have already. In any case: 1. What happens if I reorder HDD's? You haven't really described your Storage Pool but if your data array is in a healthy state, the order may change with no negative effect. 2. Where does DSM store its configurations? DSM is stored on the first partition of each disk you install in the system (there are at least three, OS, swap and data for the storage pool). The OS and swap partitions are RAID1's of however many disks you have. You only need one to boot DSM. A disk which has non-current OS partitions is labeled "System Partition Failed" and clicking the Fix button just resyncs the RAID1 arrays to that drive. You can physically plug in a SATA drive anytime with no negative effect. I have a J4105-ITX and it's connected to a hotswap backplane, works fine.

DSM tries to use the first enumerated drive for boot (the first drive on the first controller). If that doesn't match the loader, you crash. I assume you are unplugging SSD in position #1 and plugging the 3TB disk. This is the issue. Why don't you plug the third drive when the NAS is running instead of when it is offline? Or move one of the first two 8TB HDD's to the first position before plugging the 3TB drive?

This would require a virtual USB disk, which is not a feature in ESXi. This is why the loader is a disk image - because there is no way to connect it via USB. Just buy a $20 pluggable dock, put your drive in it and pass it through.

6.1 to 6.2 is a major upgrade, not a minor one. If you were able to connect via SSH, your network is functional. You ought to be able to reboot and immediately reconnect via SSH. Perhaps you did not completely delete the .xpenoboot folder.