http://bugzilla.opensuse.org/show_bug.cgi?id=1161238 http://bugzilla.opensuse.org/show_bug.cgi?id=1161238#c1 --- Comment #1 from Felix Miata --- Created attachment 827764 --> http://bugzilla.opensuse.org/attachment.cgi?id=827764&action=edit Working 5.3.8 initrd -- You are receiving this mail because: You are on the CC list for the bug.

http://bugzilla.opensuse.org/show_bug.cgi?id=1161238 http://bugzilla.opensuse.org/show_bug.cgi?id=1161238#c3 --- Comment #3 from Felix Miata --- Created attachment 827766 --> http://bugzilla.opensuse.org/attachment.cgi?id=827766&action=edit Non-booting 5.3.18 initrd part 1 made with split -n2 The whole initrd is 1857kb too big too attach. The 5.3.18 initrd was created with add_dracutmodules+=debug uncommented in /etc/dracut.conf.d/99-debug.conf -- You are receiving this mail because: You are on the CC list for the bug.

http://bugzilla.opensuse.org/show_bug.cgi?id=1161238 http://bugzilla.opensuse.org/show_bug.cgi?id=1161238#c7 --- Comment #7 from Felix Miata --- (In reply to Takashi Iwai from comment #5)

Is this a kernel problem or a problem of the rest (e.g. dracut)? I have a virtually identical PC that has no such problem, so I seriously doubt it is a kernel fault.

Keeping needinfo until I can respond to Daniel. -- You are receiving this mail because: You are on the CC list for the bug.

http://bugzilla.opensuse.org/show_bug.cgi?id=1161238 http://bugzilla.opensuse.org/show_bug.cgi?id=1161238#c9 Daniel Molkentin changed: What |Removed |Added ---------------------------------------------------------------------------- Flags|needinfo?(mrmazda@earthlink | |.net), | |needinfo?(daniel.molkentin@ | |suse.com) | --- Comment #9 from Daniel Molkentin --- Try adding x-systemd.device-timeout=1ms to the options field in fstab. -- You are receiving this mail because: You are on the CC list for the bug.

http://bugzilla.opensuse.org/show_bug.cgi?id=1161238 http://bugzilla.opensuse.org/show_bug.cgi?id=1161238#c12 Felix Miata changed: What |Removed |Added ---------------------------------------------------------------------------- Flags|needinfo?(mrmazda@earthlink |needinfo?(thomas.blume@suse |.net) |.com), | |needinfo?(daniel.molkentin@ | |suse.com) --- Comment #12 from Felix Miata --- Created attachment 827950 --> http://bugzilla.opensuse.org/attachment.cgi?id=827950&action=edit journal from booting kernel 5.3.8 with debug and rd.debug (In reply to Thomas Blume from comment #11)

"Reached target Remote File Systems" is pretty strange, system root is normally below local-fs.target. Even more, since your nvme devices are for sure local and there is no need to set up the network in the initrd.

Could that be from installed nfs-kernel-server? All remote filesystems in /etc/fstab are noauto.

Could you capture a full boot log with the additional boot parameter: debug?

There is no /var/log/boot.log. Is boot log an alias for some portion of the journal? -- You are receiving this mail because: You are on the CC list for the bug.

http://bugzilla.opensuse.org/show_bug.cgi?id=1161238 http://bugzilla.opensuse.org/show_bug.cgi?id=1161238#c16 Felix Miata changed: What |Removed |Added ---------------------------------------------------------------------------- Flags| |needinfo?(thomas.blume@suse | |.com) --- Comment #16 from Felix Miata --- No luck finding a step-by-step HOWTO for setting up serial console attachment between PCs with no 9-pin serial ports. Though I've often seen the serial console concept referred to, I've never had to try attaching one. I have 9-pin to 9-pin cables, but nothing to connect them to on the subject PC. I tried appending rd.retry=90, but it didn't help. Isn't there some way to configure/rebuild the initrd to force a timeout? I don't recognize anything like it in the dracut man page. 92 minutes isn't long enough to get 5.3.18 to timeout. debug and rd.debug I put in the default Grub stanza so as not to forget. -- You are receiving this mail because: You are on the CC list for the bug.

http://bugzilla.opensuse.org/show_bug.cgi?id=1161238 http://bugzilla.opensuse.org/show_bug.cgi?id=1161238#c19 --- Comment #19 from Felix Miata --- Created attachment 828159 --> http://bugzilla.opensuse.org/attachment.cgi?id=828159&action=edit rdsosreport.txt booted 5.3.18 using rd.break=pre-mount, debug and rd.debug Break would not exit to shell if I used rd.break=mount. I was not able to use a vfat USB stick to copy to. Without nvme access, the system has no storage, unless USB storage is attached. -- You are receiving this mail because: You are on the CC list for the bug.

http://bugzilla.opensuse.org/show_bug.cgi?id=1161238 http://bugzilla.opensuse.org/show_bug.cgi?id=1161238#c21 Felix Miata changed: What |Removed |Added ---------------------------------------------------------------------------- Attachment #828158|0 |1 is obsolete| | --- Comment #21 from Felix Miata --- Created attachment 828167 --> http://bugzilla.opensuse.org/attachment.cgi?id=828167&action=edit journalctl -b booted 5.3.18 using rd.timeout=180, debug and rd.debug -- You are receiving this mail because: You are on the CC list for the bug.

http://bugzilla.opensuse.org/show_bug.cgi?id=1161238 http://bugzilla.opensuse.org/show_bug.cgi?id=1161238#c24 --- Comment #24 from Felix Miata --- I have two Kaby Lake systems with Intel B250 chipsets. Both are configured very similarly as to NVME, main difference being gb250 also has a pair of 1TB MBR Seagates in RAID1. This bug is about the Asus, host ab250. The other is a Gigabyte, host gb250. I took the ZTC NVME stick out of of ab250 and put in it the Mushkin NVME stick from gb250. It boots right into 15.2's 5.3.16 normally, same as when booting ab250's 15.1 from its own ZTC NVME, as it does with Tumbleweed, Fedora 31, Tubuntu 18.04, Mint 19.3 and Debian 10. I put the ZTC NVME from ab250 into gb250, and it too hangs not finding root device with both 5.3.16 and 5.3.18, but not 5.3.8. I cloned 15.2 from gb250 stick partition 11 to ab250 stick partition 16, adjusted fstab and LABEL and tried to boot from it, and it too hangs not finding root device. So, I replaced the sticks in their normal PCs, copied the 15.2 5.3.16 initrd from gb250 to ab250, and it too hangs not finding root device. I just upgraded TW on ab250 to 20200121 with new kernel 5.4.12. No hang, multi-user target reached in 4.344s. Next up I think will be a clone of 15.1 from partition 9 to partition 16, then zypper upgrade to 15.2 on 16, then if same problem, a fresh 15.2 installation to ab250's partition 16. -- You are receiving this mail because: You are on the CC list for the bug.

http://bugzilla.opensuse.org/show_bug.cgi?id=1161238 http://bugzilla.opensuse.org/show_bug.cgi?id=1161238#c28 --- Comment #28 from Felix Miata --- I allowed it to rebuild only 5.3.16. It didn't change anything that I could tell. -- You are receiving this mail because: You are on the CC list for the bug.

http://bugzilla.opensuse.org/show_bug.cgi?id=1161238 http://bugzilla.opensuse.org/show_bug.cgi?id=1161238#c30 --- Comment #30 from Felix Miata --- Created attachment 828247 --> http://bugzilla.opensuse.org/attachment.cgi?id=828247&action=edit journalctl -b attempting boot 5.3.18 after upgrading nvme0n1p16's 15.2 to dracut-049+git118.a6090e2f -- You are receiving this mail because: You are on the CC list for the bug.

http://bugzilla.opensuse.org/show_bug.cgi?id=1161238 http://bugzilla.opensuse.org/show_bug.cgi?id=1161238#c32 --- Comment #32 from Felix Miata --- 30 days later, problem remains after fresh zypper dup with 5.3.18-lp152.5.3.x86_64 and initrd rebuilds of 5.3.18-lp152.1.3 and 5.3.18-lp152.3.2. Boot from 5.3.8 and TW's 5.4.12 copied from nvme0n1p7 work normally. -- You are receiving this mail because: You are on the CC list for the bug.

http://bugzilla.opensuse.org/show_bug.cgi?id=1161238 http://bugzilla.opensuse.org/show_bug.cgi?id=1161238#c37 Felix Miata (offline until ???) changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution|--- |FIXED --- Comment #37 from Felix Miata (offline until ???) --- (In reply to Takashi Iwai from comment #36)

likely already addressed in the latest Leap 15.2 kernel.

I looked at the changelog for kernel-default-5.3.18-lp152.10.4, then installed kernel-default-5.3.18-lp152.12.2. Failure disappeared. "nvme nvme0: Identify Descriptors failed (2)" remained in dmesg. -- You are receiving this mail because: You are on the CC list for the bug.