Nirin

Dynamix Autofan - Is it known that this doesn't auto-start when the NAS/Array is booted up? I have to manually go into the plugin, and 'detect' and apply the fan again, for it to start working. Is there a way to automate this? A script or something maybe?>

Did anyone ever find the solution to this? I have the same issue, on the latest mover tuner plugin (correct version, in fact the only version on CA).

Might anyone know how to disable a specific PM in powertop? Automatically I mean, I know I can go into tunables and turn the individual ones on/off. The powertop documentation only refers to turned off sets of several (such as 'all PCI' or 'all ata'). I need to disable the one, as it is for the ethernet controller and -might- be the cause of my ethernet dying sporadically.

Does anyone have the SquishedMoo guides backed up or anything? Or other guide links? Their site seems to have gone down entirely since I couple days ago, and my Ghost setup isn't working correctly (worked for 24 hours and then died for no apparent reason)./

I'm curious about the reasoning for disabling turbo boost on the cpu? I understand that this will reduce the peak power draw, as the lower frequencies will draw less power. Fine. But it will also mean the cpu will need to run for longer periods of time to do the same tasks, so you're drawing less power but for longer. It would seem that you're not actually saving much power, and have the tradeoff of a much slower system. Just was curious, maybe having turbo turned on prevents sleep states or something? If it's just a matter of the voltage curve, then the best results would come from undervolting and then limiting the max boost a little bit, so you end up on the most efficient part of the voltage/boost curve (I would think).

Just to warn people - this link takes you to a site that uses one of those "your system has been infected with a virus, click here to fix" scam popups. Sure an ad blocker usually avoids them, but still. Not worth supporting scummy people with traffic.

Trying to add a parity fisk to a new build, but it keeps hanging on 0.0mb/s speeds after a while. Mar 4 15:02:21 Hex kernel: ------------[ cut here ]------------ Mar 4 15:02:21 Hex kernel: WARNING: CPU: 10 PID: 11284 at kernel/exit.c:814 do_exit+0x87/0x923 Mar 4 15:02:21 Hex kernel: Modules linked in: xt_CHECKSUM ipt_REJECT nf_reject_ipv4 ip6table_mangle ip6table_nat iptable_mangle vhost_net tun vhost vhost_iotlb tap veth xt_nat xt_tcpudp xt_conntrack xt_MASQUERADE nf_conntrack_netlink nfnetlink xfrm_user xfrm_algo iptable_nat nf_nat nf_conntrack nf_defrag_ipv6 nf_defrag_ipv4 xt_addrtype br_netfilter xfs md_mod zfs(PO) zunicode(PO) zzstd(O) zlua(O) zavl(PO) icp(PO) zcommon(PO) znvpair(PO) spl(O) tcp_diag inet_diag nct6775 nct6775_core hwmon_vid ip6table_filter ip6_tables iptable_filter ip_tables x_tables efivarfs bridge stp llc bonding tls i915 intel_rapl_msr intel_rapl_common x86_pkg_temp_thermal intel_powerclamp coretemp kvm_intel iosf_mbi drm_buddy i2c_algo_bit ttm drm_display_helper kvm drm_kms_helper drm crct10dif_pclmul crc32_pclmul crc32c_intel ghash_clmulni_intel sha512_ssse3 sha256_ssse3 sha1_ssse3 aesni_intel crypto_simd intel_gtt mei_hdcp mei_pxp agpgart i2c_i801 cryptd rapl input_leds joydev led_class intel_cstate wmi_bmof nvme syscopyarea Mar 4 15:02:21 Hex kernel: i2c_smbus intel_uncore mei_me ahci sysfillrect sysimgblt igc i2c_core nvme_core mei libahci vmd thermal fb_sys_fops fan video tpm_crb tpm_tis tpm_tis_core wmi tpm intel_pmc_core backlight acpi_tad acpi_pad button unix Mar 4 15:02:21 Hex kernel: CPU: 10 PID: 11284 Comm: unraidd0 Tainted: P D O 6.1.74-Unraid #1 Mar 4 15:02:21 Hex kernel: Hardware name: ASUS System Product Name/ROG STRIX B760-I GAMING WIFI, BIOS 1604 12/15/2023 Mar 4 15:02:21 Hex kernel: RIP: 0010:do_exit+0x87/0x923 Mar 4 15:02:21 Hex kernel: Code: 24 74 04 75 13 b8 01 00 00 00 41 89 6c 24 60 48 c1 e0 22 49 89 44 24 70 4c 89 ef e8 31 ed 80 00 48 83 bb b0 07 00 00 00 74 02 <0f> 0b 48 8b bb d8 06 00 00 e8 33 ec 80 00 48 8b 83 d0 06 00 00 83 Mar 4 15:02:21 Hex kernel: RSP: 0018:ffffc9000113fee0 EFLAGS: 00010286 Mar 4 15:02:21 Hex kernel: RAX: 0000000000000000 RBX: ffff888130e68000 RCX: 0000000000000000 Mar 4 15:02:21 Hex kernel: RDX: 0000000000000001 RSI: 0000000000002710 RDI: 00000000ffffffff Mar 4 15:02:21 Hex kernel: RBP: 0000000000000009 R08: 0000000000000000 R09: 0000000000aaaaaa Mar 4 15:02:21 Hex kernel: R10: 0000000000000001 R11: 0000000000000001 R12: ffff88812dbf3400 Mar 4 15:02:21 Hex kernel: R13: ffff888130599080 R14: 0000000000000000 R15: 0000000000000000 Mar 4 15:02:21 Hex kernel: FS: 0000000000000000(0000) GS:ffff88885f480000(0000) knlGS:0000000000000000 Mar 4 15:02:21 Hex kernel: CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 Mar 4 15:02:21 Hex kernel: CR2: 0000000000000157 CR3: 0000000151b86000 CR4: 0000000000750ee0 Mar 4 15:02:21 Hex kernel: PKRU: 55555554 Mar 4 15:02:21 Hex kernel: Call Trace: Mar 4 15:02:21 Hex kernel: <TASK> Mar 4 15:02:21 Hex kernel: ? __warn+0xab/0x122 Mar 4 15:02:21 Hex kernel: ? report_bug+0x109/0x17e Mar 4 15:02:21 Hex kernel: ? do_exit+0x87/0x923 Mar 4 15:02:21 Hex kernel: ? handle_bug+0x41/0x6f Mar 4 15:02:21 Hex kernel: ? exc_invalid_op+0x13/0x60 Mar 4 15:02:21 Hex kernel: ? asm_exc_invalid_op+0x16/0x20 Mar 4 15:02:21 Hex kernel: ? do_exit+0x87/0x923 Mar 4 15:02:21 Hex kernel: make_task_dead+0x11c/0x11c Mar 4 15:02:21 Hex kernel: rewind_stack_and_make_dead+0x17/0x17 Mar 4 15:02:21 Hex kernel: RIP: 0000:0x0 Mar 4 15:02:21 Hex kernel: Code: Unable to access opcode bytes at 0xffffffffffffffd6. Mar 4 15:02:21 Hex kernel: RSP: 0000:0000000000000000 EFLAGS: 00000000 ORIG_RAX: 0000000000000000 Mar 4 15:02:21 Hex kernel: RAX: 0000000000000000 RBX: 0000000000000000 RCX: 0000000000000000 Mar 4 15:02:21 Hex kernel: RDX: 0000000000000000 RSI: 0000000000000000 RDI: 0000000000000000 Mar 4 15:02:21 Hex kernel: RBP: 0000000000000000 R08: 0000000000000000 R09: 0000000000000000 Mar 4 15:02:21 Hex kernel: R10: 0000000000000000 R11: 0000000000000000 R12: 0000000000000000 Mar 4 15:02:21 Hex kernel: R13: 0000000000000000 R14: 0000000000000000 R15: 0000000000000000 Mar 4 15:02:21 Hex kernel: </TASK> Mar 4 15:02:21 Hex kernel: ---[ end trace 0000000000000000 ]---

Dynamix Fan Control Currently in the log it's showing that the 'hottest drive' is 48c, and it setting the speed of the fan appropriately. Fine. However the hottest drive is actually 65c (it's an NVME) which I was going to set to ignore anyway as the fan doesn't cool that drive, but it seems dynamix is ignoring it anyway (even though I haven't ticked the 'ignore' option in the settings for fan control). The hottest hard drive though... is only 40c. So it's not those that it's looking at either. The drive that is 48c is a sata ssd, which isn't in the array or mounted, and which isn't available in the 'ignore' list for control fan. So Dynamix seems to be registering and tracking the temps of drives not in the array, but only allows you to ignore drives that -are- in the array? Seems like a bug. Two bugs actually, as it's also ignoring the nvme for no reason. So it's ignoring one drive it shouldn't, and tracking one drive it shouldn't.

I'm having this same issue, where both my fans are showing up in the bottom info bar, and showing the correct tpm, but says 0% next to them. Functionally seems to be working fine, it's just a bug. Did you manage to find a fix?

Ok so I know the trash guides setup has folder structures which put TV/Movies/Downloads folders all on a single share, so that hardlinks and atomic moves work. This seems mostly to be a benefit to torrents (which I don't use much) but is also the simplest setup method so probably the one I'll go for (still considering it though, as having different shares for TV and Movies would allow for better split levels, but thats another topic). However I use fileflows to re-encode some files (depending on source codec, and existing files size etc). And I've realised that if sonar downloads and moves the TV show onto the array, and then fileflows (or tdarr or whatever, I use fileflows) re-encodes the files, it would create new files and delete the old ones. Which would move the nicely-contiguous files onto new parts of the disks, eventually scattering them all over the place (I'm talking months/years of this happening with various folders of stuff and updated files and such). I know one solution is to have the files download to one download folder, then be encoded and moved to -another- post-download folder by sonarr, which fileflows uses as a watch folder, and is then encoded and moved to the final destination. However this means sonarr loses the file (you can set it to unmonitor so it doesn't download again at least) but will mean sonarr will no longer have an up-to-date catalogue of all files. Is there a better solution? I did think that most of the issue comes from the files being re-encoded after being put on the array. Doing the encode while still in cache would avoid this. However I don't think there's any way I can force this to happen? Sure most of the time the encoded might happen within a couple hours, but sometimes maybe it will take longer (if there's a queue). And also there might be existing files on the array that need to be re-encoded to make them more efficient (though the only solution there is probably unbalance, or manual moves, I guess). Figured I'd ask the question and hope there's a simple solution! Edit: I don't mean fragmentation in the old 'run defrag drives' sense, as I think unraid deals with that anyway... I just don't know the right term for files being scattered around randomly.

Ahh I see, that makes sense. Interesting. So there's not really any point in using it over a raid0.

I want two SSDs to be merged into a single pool. I don't care about parity, and they are backed up separately. The only ways to do this seems to be either Raid0 (faster, but if one dies the data on both is lost), or having them both balanced as 'single' which (from what I can tell) is a JBOD setup. But there seems to be some ambiguity... I've seen some say the data fills up the drives like two separate volumes, and so if one dies the data on the other is still readable, while others say the spanning happens in such a way that if one drive dies, the data on the other is gone, making it no different from a raid0 (except maybe by using special recovery services). As in either case it's probably easier to recover from the backup anyway, it's mostly an academic question (I got curious and went down a rabbit hole). But if there's zero benefit to using the 'single' option then it seems Raid0 might as well be used (even though I generally steer clear of 0).

Interesting, what's the benefit of using the pci ports over the ones on the board? I hadn't ever heard there were issues with using on board sata ports..

I guess its just that using a gen5 pci-e just for 2 sata ports, which is like.... a fraction of a fraction of the bandwidth the x16 slot can handle, seems a bit of a waste. I'm sure I could find a better use for the x16 slot, such as double NVME or a GPU. However as the primary purpose is as a NAS, and I have space for 6 sata drives in the case, I will prioritise that. I'll try the alternatives first, but if that fails I'll pick up an ASM1166

Thanks, I'll keep that in mind, though it would mean also finding a pci-e to nvme card to replace the nvme slot. Probably the only real solution though.