First try was with the volume initialized by a TiVO. Kernel panic.
Reinitialized as GUID/VFAT with MacOS Disk Utility – Kernel panic
Reinitialized as GUID/XFS on the Rocky 9 Meerkat – Kernel panic.
Disk is a LaCie 2.5, likely Seagate in their USB enclosure.
Puzzled why Rocky 9 would like it but not Rocky 8 in the HL-4. Something abut hot plugging?
Not sure what to send for supporting documentation.
Dave Hamby
So the system panics when you connect the USB drive? You aren’t trying to boot from it.
What happens if you try other USB ports? What happens if you boot into Rocky 9 using a LiveCD image instead of Rocky 8?
Trying to rule out hardware issues or kennel version issues.
No, just connecting it seems to be enough. I have a couple of open port on the back of the HL4. Either one provokes disaster.
I was able to boot the machine from a Pop_OS live CD to clean up /etc/fstab earlier.
This particular disk is not in /etcfstab as I was initializing it to use as a TimeShift external volume. This particular disk is ancient – a survivor of many years duty as a TiVO spool volume. It mounted without issue on Rocky 9.
If I connect it to Meerkat (system76 Meerkat running Rocky 9) it behaves as expected. I was able to use Cockpit-storage to put a new partition table on it (it was something strange, not the usual suspects) and a new partition on it (XFS). I tried to move it to HL-4 and it kernel paniced.
At your suggestion, I kept looking and found the Rocky Linux live CD. Links on the Rocky Linux blog way down the page. Most projects lead with their live CD to attract new recruits.
Is it my imagination, or have cloud services resulted in a deterioration of non-cloud product quality? UBUNTU installers lack former polish.
Will try the Rocky Cinnamon live CD and report results later today. Still cleaning up mail and tedious monotony.
Dave
Both Rocky8 and Rocky 9 live CD’s came up as expected. The external disk mounted with both.
Leaving the Live CD mounted and pulling the external USB disk, the internal Rocky 8 hung in Disk Manager as shown.
I’m thinking that my earlier play has left something not quite right. Not sure what.
Half-tempted to reinstall Rocky 8 but don’t have a checklist for the 45Drives and 45HomeLab additions.
Is there one squirreled away in the KB on a page I’ve only partly read?
Dave
The Pop_OS boot is probably sufficient for what I was trying to test; differences between the Linux kernel Rocky 8 and 9 are based on (5.x vs 6.x), and that the physical ports were working. It seems like there’s something about the drive physically or partition-wise even though you’ve re-partitioned it.
Some other general things I can think to do or try would be;
Maybe someone else can help in interpreting and debugging a kernel panic. There are lots of articles about that online, but I’ve not had to do it.
I reviewed /etc/fstab and removed an unneeded CIFS mount that had served its purpose.
I also commented out but didn’t remove a mount by UUID
I remembered to reload-daemon!
Problem remained.
Rocky 8 LiveCD is happy
Rocky 9 LiveCD is happy
Something is going wrong in the disk manager per console log.
Not sure what to do next. Trawl logs and dmesg? for a clue?
Found dmesg.txt from a crash. A watchdog stall mounting the LaCie disk. There was only one, the ancient 1TB Passport. This is an extract from vmcore-dmesg.txt
sdf disk driver initialization is failing. sda..sdd are the Zpool volumes.
This device is happy on both LiveCDs. It mounts and responds to ls. Didn’t touch anything.
JetKVM is connected to USB to emulate keyboard and mouse. It appears earlier in the log.
[ 2726.357909] scsi host8: uas
[ 2726.358314] scsi 8:0:0:0: Direct-Access LaCie Rugged Mini USB3 153E PQ: 0 ANSI: 6
[ 2726.359160] sd 8:0:0:0: Attached scsi generic sg6 type 0
[ 2726.359296] sd 8:0:0:0: [sdf] Spinning up disk…
[ 2727.376160] …ready
[ 2731.498851] sd 8:0:0:0: [sdf] 3907029168 512-byte logical blocks: (2.00 TB/1.82 TiB)
[ 2731.498858] sd 8:0:0:0: [sdf] 4096-byte physical blocks
[ 2731.498944] sd 8:0:0:0: [sdf] Write Protect is off
[ 2731.498946] sd 8:0:0:0: [sdf] Mode Sense: 43 00 00 00
[ 2731.499072] sd 8:0:0:0: [sdf] Write cache: enabled, read cache: enabled, doesn’t support DPO or FUA
[ 2751.785605] NMI watchdog: Watchdog detected hard LOCKUP on cpu 9Modules linked in: rfcomm xt_CHECKSUM xt_conntrack ipt_REJECT nf_nat_tftp nf_conntrack_tftp xt_multiport xt_nat xt_addrtype xt_mark ipt_MASQUERADE xt_comment nft_compat nft_counter veth bridge stp llc overlay nft_objref nf_conntrack_netbios_ns nf_conntrack_broadcast nft_fib_inet nft_fib_ipv4 nft_fib_ipv6 nft_fib nft_reject_inet nf_reject_ipv4 nf_reject_ipv6 nft_reject nft_ct nf_tables_set nft_chain_nat nf_nat nf_conntrack nf_defrag_ipv6 nf_defrag_ipv4 ip_set nf_tables nfnetlink cmac bnep sunrpc vfat fat intel_rapl_msr intel_rapl_common wmi_bmof snd_hda_codec_realtek snd_hda_codec_generic edac_mce_amd ledtrig_audio snd_hda_codec_hdmi snd_hda_intel btusb snd_intel_dspcfg btrtl btbcm btintel snd_intel_sdw_acpi btmtk snd_hda_codec kvm bluetooth snd_hda_core irqbypass snd_hwdep snd_seq crct10dif_pclmul crc32_pclmul joydev ghash_clmulni_intel snd_seq_device sp5100_tco ecdh_generic snd_pcm pcspkr rapl rfkill i2c_piix4 k10temp
[ 2751.785640] snd_timer snd soundcore ccp wmi gpio_amdpt gpio_generic xfs libcrc32c sd_mod amdgpu sr_mod cdrom sg drm_ttm_helper ttm iommu_v2 gpu_sched i2c_algo_bit drm_buddy drm_display_helper drm_kms_helper syscopyarea sysfillrect ahci sysimgblt libahci drm r8169 nvme crc32c_intel libata nvme_core realtek t10_pi video uas usb_storage zfs(POE) dm_mirror dm_region_hash dm_log dm_mod zunicode(POE) zzstd(OE) zlua(OE) zavl(POE) icp(POE) zcommon(POE) znvpair(POE) spl(OE) fuse
[ 2751.785660] CPU: 9 PID: 0 Comm: swapper/9 Kdump: loaded Tainted: P OE -------- - - 4.18.0-553.44.1.el8_10.x86_64 #1
Hey Dave,
If you decided you wanted to go with a reinstall we have this KB article to reinstall Houston UI
This is a little bit older of a guide so it will download the old zfs module but since the new one is giving people some trouble on occasion that should be fine.
There are also some extra functions such as the identities and scheduler modules up for download that are not in the guide which can be found on our github
I hope this helps.
The 45Drives Knowledge Base is my friend. I did rebuild my Roon Server (system76 Meerkat). It was running Rock, an appliance version of Linux from Roon.app. I didn’t like the absence of backup tools in Rock. My Rock actually stores music including the odd LP transfer to digital. Most recent tinkering has been to get the backups working.
So I put Rocky 8 on Meerkat. And followed the KB for Houston UI. If I redo HL-4 it appears I have to do two KP procedures, first to add ZFS, then to add Houston UI. Are there more I missed? The ports are wrong in the Roon KB. They changed them and didn’t update the Linux installation procedure.
That old 1 TB Passport dates back to 2011 or so and served as a TiVO spool volume for most of those years. Since 2017 or so, it has been idle.
Rocky (HL-4 named after Rocky greyhound) may have a bit of a ding. Running the internal OS, it is still colicky when a foreign disk is connected. If I boot any LiveCD, the ancient Passport hot plugs just fine. Nothing on it at the moment.
They don’t like APFS none! Meerkat (the Roon host) could care less. It just tells you it doesn’t know whatever filesystem is on the foreign disk.
This looks self-inflicted but is also livable for now. I imagine it will go away after an update.
Thanks,
