Daniel15

Thank you for posting your settings! I'll compare mine and figure out what's different.

The mATX board. It's possible I'm just doing something wrong.

I upgraded. Working fine except for the iGPU. For some reason, it's not in the lspci output, only the IPMI/BMC one is there. I've tried changing the primary display in the UEFI but so far it hasn't helped. Because of this, I had to disable the Intel SR-IOV plugin and stop using the iGPU in VMs. I'll try figure it out.

My knowledge was a bit outdated. Making the changes to the extra parameters should still work, but you can actually do it in the UI now without going to advanced mode: Select "Container" for network type, then the right container for "Container Network" If that still doesn't work, please provide the error message

That log is from when I boot with the plugin enabled (seems to hang), and also as soon as I enable the plugin (system still works fine, but the plugin installation hangs). Diagnostics attached.: danhome-diagnostics-20250111-1254.zip I'm going to upgrade to Unraid 7 and see if that helps at all.

Turns out my issue was the same as this known issue:

This is easy - For each Docker container you want to route through it, set the network type to "None", enable the "advanced view" at the top right, then add: --net=container:gluetun to the extra parameters field, where "gluetun" is the name of the Docker container to route through. I'm not sure it's possible to route all server traffic through the VPN with Unraid 6. Unraid 7 adds a new feature where you can add Docker containers to the tailnet (https://docs.unraid.net/unraid-os/manual/security/tailscale/#adding-tailscale-to-docker-containers). This lets any devices in your tailnet route through the VPN Docker container as a Tailscale exit node. Unraid 6 doesn't have this feature. Even with that feature, I'm not sure if you could route all the server's traffic through it.

This is a good plugin. I've been using it for a while. Thanks for working on it! I wish this functionality was built-in to Unraid. Seems like a good fit given Docker and KVM are both built-in.

I figured out that there's an "Interface Rules" section in the network config that can bind particular interfaces to particular names, but it only appears if there's multiple NICs in the system. The two onboard NICs were bound to VFIO, so there was only one NIC in use with Unraid, and so I didn't see this section. I was able to unbind the two in-built NICs from VFIO via "System Devices", reboot, update the Interface Rules to set the 10Gbps NIC's MAC to "eth1", re-bind the built-in NICs to VFIO, reboot again, and the 10Gbps NIC remained as eth1 on the next reboot. I'll see if it works long-term.

My 10Gbps network card keeps switching between eth0 and eth1 between reboots. Is there a way to enable predictable network interfaces names (the ensXXXXX names instead of the legacy ethX ones) with Unraid, like what other Linux distros do by default?

Unraid doesn't use .NET, so this general protection fault must have come from a different app you're running on your server. Likely one of the *arr apps (Lidarr, Radarr, etc) since they're all .NET apps. The attached syslog doesn't show a general protection fault, but it does have a "BUG: unable to handle page fault" which is usually indicative of either memory or disk swap issues (if you have swap enabled), but you said you ran memtest already. Hmm.

I'm trying to upgrade from Unraid 6.12.4 to 6.12.14 but I'm getting a general protection fault when the SR-IOV plugin is installed. I've tried deleting it and reinstalling it, too. I have to start in safe mode and delete the plugin to get the system to properly boot. Any suggestions? Jan 5 19:32:49 DanHome kernel: Setting dangerous option enable_guc - tainting kernel Jan 5 19:32:49 DanHome kernel: i915 0000:00:02.0: Running in SR-IOV PF mode Jan 5 19:32:49 DanHome kernel: i915 0000:00:02.0: [drm] VT-d active for gfx access Jan 5 19:32:49 DanHome kernel: general protection fault, probably for non-canonical address 0x783230252f7b1f70: 0000 [#1] PREEMPT SMP NOPTI Jan 5 19:32:49 DanHome kernel: CPU: 2 PID: 19823 Comm: modprobe Tainted: P U O 6.1.118-Unraid #1 Jan 5 19:32:49 DanHome kernel: Hardware name: ASUSTeK COMPUTER INC. System Product Name/Pro WS W680M-ACE SE, BIOS 3602 05/24/2024 Jan 5 19:32:49 DanHome kernel: RIP: 0010:osq_lock+0x59/0xdb Jan 5 19:32:49 DanHome kernel: Code: dd 89 55 10 48 89 4d 00 89 45 14 41 87 04 24 85 c0 75 07 b0 01 e9 89 00 00 00 ff c8 48 98 48 03 1c c5 60 cb 16 82 48 89 5d 08 <48> 89 2b 65 48 8b 14 25 80 cb 01 00 8b 45 10 85 c0 75 d6 48 8b 02 Jan 5 19:32:49 DanHome kernel: RSP: 0018:ffffc90028837a50 EFLAGS: 00010202 Jan 5 19:32:49 DanHome kernel: RAX: ffffffffffff8880 RBX: 783230252f7b1f70 RCX: 0000000000000000 Jan 5 19:32:49 DanHome kernel: RDX: 0000000000000000 RSI: ffffffff820dabd1 RDI: ffffffff820db0e6 Jan 5 19:32:49 DanHome kernel: RBP: ffff889fff4aed40 R08: ffff88819e414000 R09: ffffffff829583f0 Jan 5 19:32:49 DanHome kernel: R10: 00003fffffffffff R11: fefefefefefefeff R12: ffff88810867349c Jan 5 19:32:49 DanHome kernel: R13: ffff888102072000 R14: ffff888102072000 R15: 0000000000000002 Jan 5 19:32:49 DanHome kernel: FS: 0000150688b9b740(0000) GS:ffff889fff480000(0000) knlGS:0000000000000000 Jan 5 19:32:49 DanHome kernel: CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 Jan 5 19:32:49 DanHome kernel: CR2: 0000150687fd8000 CR3: 00000001a6ee4000 CR4: 0000000000750ee0 Jan 5 19:32:49 DanHome kernel: PKRU: 55555554 Jan 5 19:32:49 DanHome kernel: Call Trace: Jan 5 19:32:49 DanHome kernel: <TASK> Jan 5 19:32:49 DanHome kernel: ? __die_body+0x1a/0x5c Jan 5 19:32:49 DanHome kernel: ? die_addr+0x38/0x51 Jan 5 19:32:49 DanHome kernel: ? exc_general_protection+0x30f/0x345 Jan 5 19:32:49 DanHome kernel: ? asm_exc_general_protection+0x22/0x30 Jan 5 19:32:49 DanHome kernel: ? osq_lock+0x59/0xdb Jan 5 19:32:49 DanHome kernel: __mutex_lock.constprop.0+0x74/0x2eb Jan 5 19:32:49 DanHome kernel: ? _dev_info+0x6c/0x87 Jan 5 19:32:49 DanHome kernel: device_del+0x37/0x31d Jan 5 19:32:49 DanHome kernel: ? i915_ggtt_probe_hw+0x6c4/0x6ef [i915] Jan 5 19:32:49 DanHome kernel: platform_device_del+0x21/0x70 Jan 5 19:32:49 DanHome kernel: platform_device_unregister+0xf/0x19 Jan 5 19:32:49 DanHome kernel: sysfb_disable+0x2b/0x54 Jan 5 19:32:49 DanHome kernel: aperture_remove_conflicting_pci_devices+0x1e/0x82 Jan 5 19:32:49 DanHome kernel: i915_driver_probe+0x6b3/0xb35 [i915] Jan 5 19:32:49 DanHome kernel: ? slab_free_freelist_hook.constprop.0+0x3b/0xaf Jan 5 19:32:49 DanHome kernel: local_pci_probe+0x3d/0x81 Jan 5 19:32:49 DanHome kernel: pci_device_probe+0x18f/0x1e3 Jan 5 19:32:49 DanHome kernel: ? sysfs_do_create_link_sd+0x71/0xb7 Jan 5 19:32:49 DanHome kernel: really_probe+0x115/0x282 Jan 5 19:32:49 DanHome kernel: __driver_probe_device+0xc0/0xf2 Jan 5 19:32:49 DanHome kernel: driver_probe_device+0x1f/0x77 Jan 5 19:32:49 DanHome kernel: ? __device_attach_driver+0x97/0x97 Jan 5 19:32:49 DanHome kernel: __driver_attach+0xd7/0xee Jan 5 19:32:49 DanHome kernel: ? __device_attach_driver+0x97/0x97 Jan 5 19:32:49 DanHome kernel: bus_for_each_dev+0x6e/0xa7 Jan 5 19:32:49 DanHome kernel: bus_add_driver+0xd8/0x1d0 Jan 5 19:32:49 DanHome kernel: driver_register+0x99/0xd7 Jan 5 19:32:49 DanHome kernel: i915_init+0x1d/0x80 [i915] Jan 5 19:32:49 DanHome kernel: ? 0xffffffffa1431000 Jan 5 19:32:49 DanHome kernel: do_one_initcall+0x82/0x19f Jan 5 19:32:49 DanHome kernel: ? kmalloc_trace+0x43/0x52 Jan 5 19:32:49 DanHome kernel: do_init_module+0x4b/0x1d4 Jan 5 19:32:49 DanHome kernel: __do_sys_init_module+0xb6/0xf9 Jan 5 19:32:49 DanHome kernel: do_syscall_64+0x65/0x7b Jan 5 19:32:49 DanHome kernel: entry_SYSCALL_64_after_hwframe+0x6e/0xd8 Jan 5 19:32:49 DanHome kernel: RIP: 0033:0x150688cbadfa Jan 5 19:32:49 DanHome kernel: Code: 48 8b 0d 21 20 0d 00 f7 d8 64 89 01 48 83 c8 ff c3 66 2e 0f 1f 84 00 00 00 00 00 0f 1f 44 00 00 49 89 ca b8 af 00 00 00 0f 05 <48> 3d 01 f0 ff ff 73 01 c3 48 8b 0d ee 1f 0d 00 f7 d8 64 89 01 48 Jan 5 19:32:49 DanHome kernel: RSP: 002b:00007ffc0b17f458 EFLAGS: 00000246 ORIG_RAX: 00000000000000af Jan 5 19:32:49 DanHome kernel: RAX: ffffffffffffffda RBX: 00000000004293b0 RCX: 0000150688cbadfa Jan 5 19:32:49 DanHome kernel: RDX: 000000000042f640 RSI: 00000000005244b8 RDI: 0000150687ab4010 Jan 5 19:32:49 DanHome kernel: RBP: 000000000042f640 R08: 0000000000000007 R09: 000000000042fa30 Jan 5 19:32:49 DanHome kernel: R10: 0000000000000005 R11: 0000000000000246 R12: 0000150687ab4010 Jan 5 19:32:49 DanHome kernel: R13: 0000000000000016 R14: 00000000004294e0 R15: 000000000042f640 Jan 5 19:32:49 DanHome kernel: </TASK> Jan 5 19:32:49 DanHome kernel: Modules linked in: i915(O+) drm_buddy drm_display_helper intel_gtt vhost_net vhost tap kvm_intel kvm xt_CHECKSUM ipt_REJECT nf_reject_ipv4 ip6table_mangle iptable_mangle vhost_iotlb wireguard curve25519_x86_64 libcurve25519_generic libchacha20poly1305 chacha_x86_64 poly1305_x86_64 ip6_udp_tunnel udp_tunnel libchacha macvlan nft_compat xt_nat veth xt_conntrack nf_conntrack_netlink xfrm_user xfrm_algo xt_addrtype br_netfilter xt_MASQUERADE xt_tcpudp xt_mark xfs tun nf_tables nfnetlink ip6table_nat iptable_nat nf_nat nf_conntrack nf_defrag_ipv6 nf_defrag_ipv4 nfsd auth_rpcgss oid_registry lockd grace sunrpc md_mod tcp_diag inet_diag ipmi_devintf nct6775 nct6775_core hwmon_vid ip6table_filter ip6_tables iptable_filter ip_tables x_tables efivarfs 8021q garp mrp bridge stp llc zfs(PO) intel_rapl_msr intel_rapl_common iosf_mbi x86_pkg_temp_thermal intel_powerclamp coretemp zunicode(PO) zzstd(O) zlua(O) zavl(PO) ast crct10dif_pclmul icp(PO) drm_vram_helper crc32_pclmul crc32c_intel Jan 5 19:32:49 DanHome kernel: i2c_algo_bit ghash_clmulni_intel drm_ttm_helper sha512_ssse3 ttm sha256_ssse3 sha1_ssse3 drm_kms_helper aesni_intel crypto_simd cryptd drm zcommon(PO) mei_pxp znvpair(PO) mei_hdcp rapl agpgart ipmi_ssif syscopyarea intel_cstate spl(O) wmi_bmof i2c_i801 tpm_crb sr_mod sysfillrect mei_me cdc_ether i2c_smbus tpm_tis sysimgblt nvme ahci intel_uncore acpi_ipmi input_leds i2c_core fb_sys_fops usbnet video tpm_tis_core igc atlantic mei libahci nvme_core cdrom joydev led_class mii vmd thermal wmi fan ipmi_si tpm backlight acpi_pad intel_pmc_core acpi_tad button unix [last unloaded: kvm] Jan 5 19:32:49 DanHome kernel: ---[ end trace 0000000000000000 ]--- Jan 5 19:32:49 DanHome kernel: RIP: 0010:osq_lock+0x59/0xdb Jan 5 19:32:49 DanHome kernel: Code: dd 89 55 10 48 89 4d 00 89 45 14 41 87 04 24 85 c0 75 07 b0 01 e9 89 00 00 00 ff c8 48 98 48 03 1c c5 60 cb 16 82 48 89 5d 08 <48> 89 2b 65 48 8b 14 25 80 cb 01 00 8b 45 10 85 c0 75 d6 48 8b 02 Jan 5 19:32:49 DanHome kernel: RSP: 0018:ffffc90028837a50 EFLAGS: 00010202 Jan 5 19:32:49 DanHome kernel: RAX: ffffffffffff8880 RBX: 783230252f7b1f70 RCX: 0000000000000000 Jan 5 19:32:49 DanHome kernel: RDX: 0000000000000000 RSI: ffffffff820dabd1 RDI: ffffffff820db0e6 Jan 5 19:32:49 DanHome kernel: RBP: ffff889fff4aed40 R08: ffff88819e414000 R09: ffffffff829583f0 Jan 5 19:32:49 DanHome kernel: R10: 00003fffffffffff R11: fefefefefefefeff R12: ffff88810867349c Jan 5 19:32:49 DanHome kernel: R13: ffff888102072000 R14: ffff888102072000 R15: 0000000000000002 Jan 5 19:32:49 DanHome kernel: FS: 0000150688b9b740(0000) GS:ffff889fff480000(0000) knlGS:0000000000000000 Jan 5 19:32:49 DanHome kernel: CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 Jan 5 19:32:49 DanHome kernel: CR2: 0000150687fd8000 CR3: 00000001a6ee4000 CR4: 0000000000750ee0 Jan 5 19:32:49 DanHome kernel: PKRU: 55555554 Jan 5 19:32:49 DanHome kernel: note: modprobe[19823] exited with preempt_count 1

If you want VNC to keep working too (eg for emergency access when the network config is broken for whatever reason), you'll have to add a second graphics card to the VM and set it to "Virtual". There's a plus button to add another graphics card.

That CPU is a decade old, so I'm not surprised. A lot has changed in ~10 generations of CPUs, to the point where even a modern low-end i3 would perform better than that dated E5. The significantly faster RAM definitely helps too That's weird... Was it new or second-hand? If new then I'd return it and get a replacement.

I have the MicroATX version of the board, which has embedded IPMI instead of needing a PCIe card, so it's okay for the x1 slot to be covered (I'm not using it for anything else). I have a Noctua NH-D15 on mine, in a case that it's not even supposed to fit in - a Fractal Design Node 804. Very tight fit, but it does fit with maybe 5-10mm of clearance from the top of the heatsink to the side panel

LSI 9207-8i is a PCIe 3.0 x8 card. It'll work in a 4x slot, but the maximum bandwidth will be halved, which may cause a bottleneck when you're using a lot of the drives at the same time. You could bifurcate one of the PCIe 5.0 x16 slots so it essentially becomes two x8 slots or four x4 slots, but I don't have any experience doing that so I don't know how to do it. PCIe 4.0 x4 has the same bandwidth as PCIe 3.0 x8, so if you can find an equivalent card in a PCIe 4.0 version, it'll work in a 4.0 or 5.0 x4 slot.

What kernel version does beta3 use, and which version did the previous beta use?

If you're reinitializing the card, you also need to re-enable SR-IOV: echo 7 > /sys/devices/pci0000:00/0000:00:02.0/sriov_numvfs Before doing that, double-check the address (00:02:0) in lspci output and adjust it if needed.

With the Xe driver (which won't need any plugin), or with the i915 driver (which is the one this plugin is for)? The conversation above was about the Xe driver specifically.

They said "newer platforms will be supported later".

Thanks for the reply. I got it working with your template. Since I'm using a Docker bridge for everything, I needed to change the Meili URL to http://meilisearch:7700 and the browser address to ws://browserless-v2:3000?token=..... (host names as the names of the Docker containers) to get them to work properly. Looks like it's working great though! Now I need to try the OpenAI integration and import thousands of bookmarks from Pinboard

What's better - Using this Docker image, or using Docker Compose Manager with upstream's docker-compose.yml file? Docker Compose seems like it'd be easier to maintain since it's the "official" way, but regular Docker integrates better with Unraid...

SR-IOV lets you split the iGPU up to seven ways. Either the host (for Docker) + 6 VMs, or 7 VMs if you don't use the GPU on the host.

In theory it should "just work" once the new Intel drivers are ready for production use, assuming the new drivers support your Intel GPU. This is from Intel's announcement: Those platforms are Tiger Lake, Alder Lake, Meteor Lake, Arctic Sound M, and Ponte Vecchio respectively.

Sometimes some of the onboard features share/reuse PCIe lines, and thus some PCIe or NVMe slots stop working when you use them. It's weird to see all of them stop working though. The frustrating thing with this board is that Asus refuse to release a block diagram for it, which would allow us to see all the shared PCIe lanes and what they're attached to. I emailed them asking for one and they said they couldn't release it publicly <_<