WiteWulf

@imdgg as Orphée pointed, there's an ongoing issue with databases and/or docker crashing the DS3615xs image, both on baremetal and virtualised platforms, that's not present on 918+. However that may not be the root of your problem. Your logs are full of the following message: 2021-09-25T22:31:50+08:00 DS3615xs synostgd-cache[9725]: cache_monitor.c:2074 [Error] Can't support DS with cpu number (1) At a guess your ESXi guest only has one vCPU configured? Try increasing it to 2 and see how you get on. You've got 4 cores on that CPU (8 with hyperthreading), show it some love 😁

As Haydibe pointed out, it's not actually using the kernel source at the moment, it's using the toolkit instead, so the kernel source line is actually irrelevant.

No, you're missing my point: just having a compatible chipset does not guarantee that a device will work. This is the same for any device on most operating systems. The device (NIC, bluetooth adapter, wifi adapter, whatever), identifies itself to the OS with a hardware ID. If it's PCIe card as in this instance it will use a PCI ID. The operating system then looks up that PCI ID in all the drivers it has loaded to see if there is anything that matches. It has no idea what chips are in the device up until this point, what the device does, nor what driver is required. If it finds a driver with a matching PCI ID it will then try to access the device using that driver. The problem is that hardware manufacturers sometimes change the hardware ID of their devices over their lifetime, so you could have two NICs that look identical, with the same chipset, but one will match the driver's ID list and one will not. This is particularly prevalent in hardware that is primarily developed for Windows use, but open source drivers are developed for linux. That is why the list of matching PCI IDs was published in that thread previously, as some Intel NICs work with DSM and some do not, although look identical.

Of course, thanks for pointing both of those out! I've not had enough coffee today and am still not getting my head around virtualising DSM 😁

Interesting idea, do those PCIe NVMe adapters work on xpenlogy, then? This sort of thing, for example: https://www.amazon.co.uk/SupaGeek-PCIe-Express-Adapter-Card/dp/B07CBJ6RH7/ref=pd_lpo_3?pd_rd_i=B07CBJ6RH7&psc=1

Ah, I hadn't thought of that! I'd assumed that it was on a different controller, actually. Raw disk is definitely the only option, then, unless I get PCIe HBA and get redpill/DSM7 working with the internal NIC (which plenty of people seem to be doing now).

Not all Intel NICs work on DSM. It's not just the chipset that needs to be supported, the PCI ID of the card also has to match a list that linux uses to decide what driver to use. More info here: Now that's obviously from an older version of DSM, but I very much doubt it's changed.

As I mentioned: all four HDDs are currently full with DSM data. I can't resize them so I need to pass those through to the VM (either as raw disks or via the controller). My plan is to use an SSD on the ODD connector inside the Gen8 for datastore. There's one in there already that's configured as a read-cache for DSM, but I'm not convinced it's making a lot of difference. I've read a few places online explaining how to keep the datastore on the same USB stick you boot from, or the SD card slot on the motherboard, but I'm loathe to run from flash storage.

I'm contemplating migrating moving my baremetal install on an HP Gen8 Microserver to ESXi (ESXi because I use it at work and am more familiar with it than Proxmox). It seems pretty simple: just replace the xpenology USB boot stick I'm currently using with an ESXi boot stick, create a VM for DSM with a virtual boot image, pass through the existing disks and boot it up. DSM will do the "I've detected disks from another server, do you want to migrate?" thing, and I'm done, right? My main question before I do this is: given that I'm running the SATA controller on the Gen8 in AHCI mode (ie. no "hardware" RAID), should I pass through the controller to the VM, or the individual disks in Raw Disk Mode? Is there any performance benefit to either? The disks (4x3TB) are full with DSM data, obviously, so I'll not be able to use that set of disks for any other ESXi guests, but I'm considering getting an HBA at some point to add some extra storage.

Since moving to the latest redpill code (3474d9b) I'm finding my system less stable than it was before with respect to docker containers. My influxdb container would crash the system every time I started it, but the others I had were typically stable and non-problematic. I'm now seeing immediate kernel panics when starting a mysql container that previously didn't cause any problems. I've gone back to a slightly older build (021ed51), and mysqld (operating as a database backend for librenms) starts and runs without problems.

My CPU (E3-1265L V2) is only supported on the bromolow image.

@ThorGroup it still kernel panics when launching an influxdb docker container with "register_pmu_shim" deleted from redpill_main.c as suggested: [ 338.055690] Kernel panic - not syncing: Watchdog detected hard LOCKUP on cpu 6 [ 338.091670] CPU: 6 PID: 21097 Comm: containerd-shim Tainted: PF O 3.10.108 #42214 [ 338.132114] Hardware name: HP ProLiant MicroServer Gen8, BIOS J06 04/04/2019 [ 338.168045] ffffffff814a2759 ffffffff814a16b1 0000000000000010 ffff880409b88d60 [ 338.205031] ffff880409b88cf8 0000000000000000 0000000000000006 0000000000000001 [ 338.241507] 0000000000000006 ffffffff80000001 0000000000000030 ffff8803f4d4dc00 [ 338.278173] Call Trace: [ 338.290006] <NMI> [<ffffffff814a2759>] ? dump_stack+0xc/0x15 [ 338.318839] [<ffffffff814a16b1>] ? panic+0xbb/0x1df [ 338.342727] [<ffffffff810a9eb8>] ? watchdog_overflow_callback+0xa8/0xb0 [ 338.375043] [<ffffffff810db7d3>] ? __perf_event_overflow+0x93/0x230 [ 338.405804] [<ffffffff810da612>] ? perf_event_update_userpage+0x12/0xf0 [ 338.438356] [<ffffffff810152a4>] ? intel_pmu_handle_irq+0x1b4/0x340 [ 338.469218] [<ffffffff814a9d06>] ? perf_event_nmi_handler+0x26/0x40 [ 338.500130] [<ffffffff814a944e>] ? do_nmi+0xfe/0x440 [ 338.525060] [<ffffffff814a8a53>] ? end_repeat_nmi+0x1e/0x7e [ 338.552408] <<EOE>> [ 338.562333] Rebooting in 3 seconds.. Tried it another time, similar crash, this time directly referncing influxdb in the output: Kernel panic - not syncing: Watchdog detected hard LOCKUP on cpu 0 [ 165.610117] CPU: 0 PID: 21435 Comm: influxd Tainted: PF O 3.10.108 #42214 [ 165.646799] Hardware name: HP ProLiant MicroServer Gen8, BIOS J06 04/04/2019 [ 165.680969] ffffffff814a2759 ffffffff814a16b1 0000000000000010 ffff880409a08d60 [ 165.717087] ffff880409a08cf8 0000000000000000 0000000000000000 0000000000000001 [ 165.753547] 0000000000000000 ffffffff80000001 0000000000000030 ffff8803f5267c00 [ 165.789827] Call Trace: [ 165.801740] <NMI> [<ffffffff814a2759>] ? dump_stack+0xc/0x15 [ 165.830687] [<ffffffff814a16b1>] ? panic+0xbb/0x1df [ 165.855249] [<ffffffff810a9eb8>] ? watchdog_overflow_callback+0xa8/0xb0 [ 165.888617] [<ffffffff810db7d3>] ? __perf_event_overflow+0x93/0x230 [ 165.919963] [<ffffffff810da612>] ? perf_event_update_userpage+0x12/0xf0 [ 165.952655] [<ffffffff810152a4>] ? intel_pmu_handle_irq+0x1b4/0x340 [ 165.983546] [<ffffffff814a9d06>] ? perf_event_nmi_handler+0x26/0x40 [ 166.013584] [<ffffffff814a944e>] ? do_nmi+0xfe/0x440 [ 166.038375] [<ffffffff814a8a53>] ? end_repeat_nmi+0x1e/0x7e [ 166.065408] <<EOE>> [ 166.075520] Rebooting in 3 seconds..

I think you're right. This should be a correct bromolow 7.0.1-RC1 stanza: { "id": "bromolow-7.0.1-42214", "platform_version": "bromolow-7.0.1-42214", "user_config_json": "bromolow_user_config.json", "docker_base_image": "debian:8-slim", "compile_with": "toolkit_dev", "redpill_lkm_make_target": "prod-v7", "downloads": { "kernel": { "url": "https://sourceforge.net/projects/dsgpl/files/Synology%20NAS%20GPL%20Source/25426branch/bromolow-source/linux-3.10.x.txz/download", "sha256": "18aecead760526d652a731121d5b8eae5d6e45087efede0da057413af0b489ed" }, "toolkit_dev": { "url": "https://sourceforge.net/projects/dsgpl/files/toolkit/DSM7.0/ds.bromolow-7.0.dev.txz/download", "sha256": "a5fbc3019ae8787988c2e64191549bfc665a5a9a4cdddb5ee44c10a48ff96cdd" } }, "redpill_lkm": { "source_url": "https://github.com/RedPill-TTG/redpill-lkm.git", "branch": "master" }, "redpill_load": { "source_url": "https://github.com/jumkey/redpill-load.git", "branch": "develop" } },

As others have suggested: this isn't right for you yet. Yes, this is gatekeeping, but there's a beta coming soon hopefully. But be aware that there will never be a simple image to download for redpill. Read the thread (particularly Thorgroup's posts) and you'll hopefully understand why.

Does anyone keep up with news from Synology themselves? Just wondering if there's any sign of 7.0.1 progressing beyond the release candidate 🤔

First post on the first page of this thread: https://github.com/RedPill-TTG/redpill-lkm

I'm just about to knock off for the day (and it's my partner's birthday, so no hacking tonight! ), but I'll give that a go tomorrow morning (UK time).

Yes, that's exactly what I observed. It was fine with a small folder of ~250 photos, but when I pointed it at a 2TB folder of music files it kernel panic'd after a few minutes.

BTW, in testing the kernel panic stuff I installed "Photos" to my machine and was pleasantly surprised to find the face detection is working 👍

Little update on the docker kernel panics thing some of us were diagnosing last week: - @Orphée and "labrous" commented on the github issue that they were seeing similar issues with the Moments app on 6.2.4 and Photos on 7.0.1 (lock ups on VMs, kernel panics on baremetal) - the common thing I'm seeing here is databases: most of the docker crashes I've seen have involved mariadb, influxdb, mysqld or elasticsearch

Redpill does not output anything to the screen after this message. All further diagnostic information is output to the serial port. @ThorGroup *please* put a message on there saying there's no more output, like Jun's does

You have 8GB internet at home 😮 I mean, my workplace has 40Gb/s, but we do 100Gb/s across the campus backbone, 20Gb/s to the distribution layer and 1Gb/s to most edge devices. My personal NAS (running Plex) hosted at work is on a teamed 2x1Gb link and the 8 disk array in that can't saturate that link. What do you *do* with 8Gb? 🤣

Sorry for going off topic for a moment, but I see a few people wanting 10GB ethernet. I wonder, do you have a 10GB ethernet switch, and any other 10GB capable devices on your LAN (or multiple 1Gb clients capable needing an aggregate 10GB throughput)? Do you even have a storage array capable of saturating a 10GB link? Just wondering what the use case is for hardware like this, as it's the sort of thing that's very rarely properly utilised outside of datacentres...

FYI, signatures aren't visible to users on mobile devices.

Yeah, it’ll identify the disks as being from another server and offer to do a migration for you. So long as you give it a pat file that matches the version and architecture you’ve built the boot image for you you’ll be fine.