[FAIL]Dual boot. Random errors during install - but Win works flawless

MonkeyBoy · 2020-03-11 03:26:54

[UPDATE I have now tried to zap the entire disk, install arch as well as solus and elementary in a plethora of permutations on the layout. Installation is still a lottery, mostly miss, few hits and most of those prone to fail over time. Windows installs and works flawless still. I am running the lenovo full hardware diagnosis over night now but to be honest I find myself at an end of my imagination and abilities when it comes to DuckDuckGo. I put my hope to divine intervention.]

Hi,
(Please remove and/or point me in the right direction if this issue has already been addressed. Though I have searched the forum as well as the interwebz I might have missed out on the obvious)

Here goes:
I have a Lenovo T480 with an ssd and win10 preinstalled. I booted arch live usb and started off as by the install guide, nothing out of the ordinary, I have done this several times before on other machines. Now it is time to mount root and that works. Some times. Next time lsblk gives that the nvme0n1 has dissappeared. Start again, now I can mount root and then I make the dirs for home and var and what have you - boom! All in a sudden everything just halts and odd message pops up, like this for example:
A screen dump. Ish. (Sorry, not tech savvy enough)

Next time I execute loadkeys and without return being hit everything freezes for quite a while, a new message like the one above pops up but this time the /var or the like is not rw.

To me it is just random. For testing purposes I have done identical procedures, following the installation guide to the t, but still error messages pops up at random places in the process. My gut feeling says: something iffy with hardware (eg. the ssd). But windows works flawless! I have emptied the entire drive, rid it of partition tables, wiped it with zeroes and probably some more things just for being desperate - same thing. Reinstall windows - works flawless, try with arch again and the same outcome. Have made another live usb made with rufus (latest iso) - same result. I have spent a day with this - What gives?!

Regards // Desperate Monkey

Last edited by MonkeyBoy (2020-04-04 03:09:35)

Wild Penguin · 2020-03-11 08:36:59

Hi,

It seems there indeed are problems in reading that NVME drive, as you have deduced. As you are getting I/O errors, creating new partitions / erasing the disk wont fix that. Neither will creating a new bootable USB. You can always do those but it is in vain, in this case.

What I would do is: re-seat the drive. Despite it working in Windows, might be just because of chance. Better back up anything on the drive you don't want to lose, just in case.

There may be several reasons why it works in Windows; for example, it's drivers may have better error handling. It may be able to downclock the speed of the link to the NVMe, and maybe it just works better at lower speeds. Check Windows logs for any error messages about the NVMe.

Then, it could also be that the Lenovo T480 just has some (proprietary) NVMe H/W implementation which does not have a correctly working driver for Linux. I believe this is a very unlikely possibility, however it is still a possibility. But if this was the case, there would be many reports of this issue (unless the Laptop you have has some exotic, rare hardware, which I believe is not the case?).

I'm not sure if the above helps, just my thoughts. Maybe it can lead to a better Google search - or you may want to wait if someone else has better ideas here.

p.s. For future reference, those error messages should be in journal, try journalctl and learn to use it to get the log snippets into files which you can move / netcat somewhere. You can also enable sshd on installation the machine to ssh it from another to facilitate copy+pasting text here. Posting pictures is not endorsed, and makes it just more difficult to those reading your post and as such you are less likely to get help.

Last edited by Wild Penguin (2020-03-11 08:38:53)

Lone_Wolf · 2020-03-11 11:56:55

https://wiki.archlinux.org/index.php/Du … th_Windows

My money is on fast startup giving problems, but make sure you read the whole page thoroughly several times.

MonkeyBoy · 2020-03-11 15:29:54

[EDIT Due to lapsus' of my own doing] [EDIT Found a possible different explanation to the problem - and solution]

@Lone_Wolf: Sorry for being unclear, the results were the same with a clean arch install, no win on the drive. Still appreciate your time and effort a lot though.

@ Wild Penguin: Re-seating the NVMe is sound advice and I choose that path. The t480 does not seem to have any out-of-the-ordinary hardware and the NVMe is a Lite-on. Though tired/stoopid/frustrated as I were yesterday I did not write down the model (bad me) but I will work it out . Anyways. I gutted my precious, unhooked the battery, unhooked the cable from the MB, unhooked the 'adapter/controller' and then removed and replaced the NVMe a couple of times to make sure there was no piece of dust or the like on the 'contact strips'. Then reverse order, starting up the engines and the NVMe was now *tadaa!* nowhere to be found. I tried a couple of times with full shut down, tried with the arch live stick but to no avail. I repeated the process from scratch one more time and then accepted facts: no go. (no shadows upon you, it was sound advice)

My guess is that the re-seating broke an already iffy connection or the like. I just hope it is not situated in the contact on the mother board because, well yeah..

EDIT: Apparently there has been a problem with the t470 to recognise the NVMe - the cable. I only found one video on it bit it is worth a try (volume levels hits the roof at 0:35 - beware!)

If the cable wont workI will proceed with buying a USB3-NVMe enclosure/adapter for ~25$ from a popular shopping site, check if the NVMe works and if it does buy an M.2 adapter from Lenovo for ~30$ and try. Otherwise buy new NVMe, new adapter from Lenovo and try. And cry if it doesn't work... Well, a select paperweight should always be an accessoire in the drawing room for the discerning gentleman (t480 carcass, anybody? Fresh out of the oven..)

Despite all I am grateful for all your help, and would hapily take any other advice and feedback on the issue, if there is. I will report back on this issue and see where we end up. Hopefully to help someone in the future if nothing else.

Great many thanks // Desperate Monkey

Reg. journal instead of photo - I tried but could not really get the error messages on the arch live stick, sometimes the error lead to a total freeze but when it wasn't I might have failed due to lack of capability/clear mind. My bad.

Last edited by MonkeyBoy (2020-03-11 18:11:55)

MonkeyBoy · 2020-03-15 16:46:17

[UPDATE]

I moved the NVMe to a USB adapter, into another computer, booted into an arch live stick and installed arch to the NVMe flawless. I.e. the NVMe seems fine and now I will buy a new cable to see if the problem sits there. Then it would be consistent withe problem of the t470.

To be continued..

MonkeyBoy · 2020-03-24 02:55:40

[UPDATE]

Tried the new cable to no avail.

Tried install another dist, 'Solus'. This worked(!), it did not show up in any boot menu but worked flawless when starting it up 'manually' (live stick, chroot). I am fine with that, I wiped the Solus partition and tried to install Arch on that single partition - it worked! I fiddled around with getting solus out of the efi-partition and get arch in place as the boot problem was still there. Alas, i could not bootrec /fixboot or rebuild bcd - I did not get any access; worked around but win said the boot files could not be copied or the like. OK, re-install win, install Arch - nope. Rid of all non-win partitions, setup partitions from scratch: root, swap, home, var, tmp. Format / -ext4, /home -ext4, /var reiserfs, /tmp reiserfs. Mount root, make dirs, mount /boot (shared with win), mounted /home - boom!

I am back at square one, this time I have some lines from journalctl for show (ix.io - brilliant!) Please find comments below

Mar 24 01:33:05 archiso kernel: audit: type=1131 audit(1585013585.583:40): pid=1 uid=0 auid=4294967295 ses=4294967295 msg='unit=systemd-timedated comm="systemd" exe="/usr/lib/systemd/systemd" hostname=? addr=? terminal=? res=success'
Mar 24 01:33:34 archiso kernel: Adding 4194300k swap on /dev/nvme0n1p7.  Priority:-2 extents:1 across:4194300k SSFS
Mar 24 01:38:44 archiso kernel: EXT4-fs (nvme0n1p10): mounted filesystem with ordered data mode. Opts: (null)
Mar 24 01:40:46 archiso kernel: EXT4-fs (nvme0n1p6): mounted filesystem with ordered data mode. Opts: (null)
Mar 24 01:41:19 archiso kernel: nvme nvme0: I/O 634 QID 4 timeout, aborting
Mar 24 01:41:19 archiso kernel: nvme nvme0: Abort status: 0x0
Mar 24 01:41:23 archiso kernel: nvme nvme0: I/O 635 QID 4 timeout, aborting
Mar 24 01:41:23 archiso kernel: nvme nvme0: Abort status: 0x0
Mar 24 01:41:50 archiso kernel: nvme nvme0: I/O 634 QID 4 timeout, reset controller
Mar 24 01:42:13 archiso kernel: nvme nvme0: Device not ready; aborting initialisation
Mar 24 01:42:13 archiso kernel: nvme nvme0: Removing after probe failure status: -19
Mar 24 01:42:13 archiso kernel: blk_update_request: I/O error, dev nvme0n1, sector 111466496 op 0x1:(WRITE) flags 0x800 phys_seg 1 prio class 0
Mar 24 01:42:13 archiso kernel: Buffer I/O error on dev nvme0n1p8, logical block 1048560, async page read
Mar 24 01:42:13 archiso kernel: Buffer I/O error on dev nvme0n1p6, logical block 557056, lost sync page write
Mar 24 01:42:13 archiso kernel: FAT-fs (nvme0n1p8): unable to read boot sector
Mar 24 01:42:13 archiso kernel: blk_update_request: I/O error, dev nvme0n1, sector 0 op 0x1:(WRITE) flags 0x800 phys_seg 0 prio class 0
Mar 24 01:42:13 archiso kernel: JBD2: Error -5 detected when updating journal superblock for nvme0n1p6-8.
Mar 24 01:42:13 archiso kernel: Aborting journal on device nvme0n1p6-8.
Mar 24 01:42:13 archiso kernel: Buffer I/O error on dev nvme0n1p6, logical block 557056, lost sync page write
Mar 24 01:42:13 archiso kernel: JBD2: Error -5 detected when updating journal superblock for nvme0n1p6-8.
Mar 24 01:42:13 archiso kernel: EXT4-fs error (device nvme0n1p6) in ext4_init_inode_table:1437: IO failure
Mar 24 01:42:13 archiso kernel: Buffer I/O error on dev nvme0n1p6, logical block 0, lost sync page write
Mar 24 01:42:13 archiso kernel: EXT4-fs (nvme0n1p6): I/O error while writing superblock
Mar 24 01:42:13 archiso kernel: nvme nvme0: failed to set APST feature (-19)

First line is timedatectl set, just to show there is no work w the disk before.
Mounted root (p10)
I made all the dirs
Mounted boot (p2) - that cannot be found in the log above(?!) but I did it, it was there in the 'prompt history', executed.
Mounted home (p6) - boom!

For ref:
p7 is swap
p8 is var
p9 is tmp

I can give it more goes if you like, the freeze/loss of nvme0n1 seems a game of chance to me so more info/log could be extricated if wanted.

I could just get Solus in place, fix the dual boot issue and be happy. But Solus makes monkey a dull boy. Arch makes happy monkey.

Grateful for all input. I want to beat this one.

Arch Linux

#1 2020-03-11 03:26:54

[FAIL]Dual boot. Random errors during install - but Win works flawless

#2 2020-03-11 08:36:59

Re: [FAIL]Dual boot. Random errors during install - but Win works flawless

#3 2020-03-11 11:56:55

Re: [FAIL]Dual boot. Random errors during install - but Win works flawless

#4 2020-03-11 15:29:54

Re: [FAIL]Dual boot. Random errors during install - but Win works flawless

#5 2020-03-15 16:46:17

Re: [FAIL]Dual boot. Random errors during install - but Win works flawless

#6 2020-03-24 02:55:40

Re: [FAIL]Dual boot. Random errors during install - but Win works flawless

Board footer