Closed grahamperrin closed 8 years ago
Interesting. Can you, or have you tried the GRUB bootloader on these machines to see if that also has the same issues? In the installer there is a dropdown that says BSD which can be changed to GRUB for testing.
… have you tried the GRUB bootloader on these machines to see if that also has the same issues? …
For the 8570p in July: yes.
For brevity, in that post I chose to not list the combinations that I had tried with each of the five notebooks but I'm sure as I can be that all possible options were tried.
https://forums.pcbsd.org/thread-20097-post-111697.html#pid111697 reminds me that 11.0-CURRENTSEPT2015 produced a bootable operating system with UEFI.
If I recall correctly, problems with installers began when automation was introduced to tell whether it was appropriate to offer MBR.
For the 8540p and 8570p cases I'm almost certain that where successful installation to GPT fails to produce a bootable OS, installation to MBR will succeed.
We may have fixed this issue for you. Now I recall Kris was having this exact issue where he could install but not boot with EFI. We found that adding to many things in loader. On certain systems we hit a memory limitation where systems would either not boot at all, or load all modules. Therefore we have been moving as much as we can to rc.conf.trueos / rc.conf (in the case of nvidia-driver).
https://github.com/trueos/trueos-core/commit/4189f91fd2d28b0be08221177323c3a9c75d50bd https://github.com/trueos/trueos-core/commit/b27a00f86358f18ec93b391c1854f2e5f3f39b19
To confirm after an install you could try to mount the zpool to /mnt using the utilities shell in the installer, and then edit /boot/loader.conf.trueos to remove the entire linux compat section as shown in the first commit. This may resolve this issue for you. Our Monday / Tuesday ISO should have the fix out of box for you to test as well.
edit /boot/loader.conf.trueos to remove the entire linux compat section
No such section here.
For what it's worth, from http://lists.pcbsd.org/pipermail/testing/2016-July/010794.html
It may be useful to compare the failure… with how an installation of Kubuntu behaves on the same notebook. …
Workaround
Not all of the space available to /dev/sda appears to be used, you can fix the GPT to use all of the space (an extra 7 blocks) or continue with the current setting?
– opt to fix the table, then restart the computer.
I performed another clean installation that was not bootable, then again allowed GParted to fix the GPT.
Before and after the fix – note the change to the last usable sector, from 976773127 to 976773134:
Welcome - Parted Magic (Linux 3.10.4-pmagic)
root@partedmagic:~# gdisk
GPT fdisk (gdisk) version 0.8.6
Type device filename, or press <Enter> to exit:
root@partedmagic:~# gdisk
GPT fdisk (gdisk) version 0.8.6
Type device filename, or press <Enter> to exit: /dev/sda
Partition table scan:
MBR: protective
BSD: not present
APM: not present
GPT: present
Found valid GPT with protective MBR; using GPT.
Command (? for help): p
Disk /dev/sda: 976773168 sectors, 465.8 GiB
Logical sector size: 512 bytes
Disk identifier (GUID): 4D7213D0-5DF2-11E6-98FC-68B599FCDB6A
Partition table holds up to 128 entries
First usable sector is 40, last usable sector is 976773127
Partitions will be aligned on 8-sector boundaries
Total free space is 30688 sectors (15.0 MiB)
Number Start (sector) End (sector) Size Code Name
1 40 204839 100.0 MiB EF00
2 204840 959965223 457.6 GiB A504
3 959965224 976742439 8.0 GiB A502
Command (? for help): q
root@partedmagic:~# date ; uname -a
Mon Aug 8 22:56:06 AKDT 2016
Linux partedmagic 3.10.4-pmagic #2 Tue Jul 30 11:06:11 CDT 2013 i686 Intel(R) Core(TM) i5 CPU M 540 @ 2.53GHz GenuineIntel GNU/Linux
root@partedmagic:~# gdisk
GPT fdisk (gdisk) version 0.8.6
Type device filename, or press <Enter> to exit: /dev/sda
Partition table scan:
MBR: protective
BSD: not present
APM: not present
GPT: present
Found valid GPT with protective MBR; using GPT.
Command (? for help): p
Disk /dev/sda: 976773168 sectors, 465.8 GiB
Logical sector size: 512 bytes
Disk identifier (GUID): 4D7213D0-5DF2-11E6-98FC-68B599FCDB6A
Partition table holds up to 128 entries
First usable sector is 40, last usable sector is 976773134
Partitions will be aligned on 8-sector boundaries
Total free space is 30695 sectors (15.0 MiB)
Number Start (sector) End (sector) Size Code Name
1 40 204839 100.0 MiB EF00
2 204840 959965223 457.6 GiB A504
3 959965224 976742439 8.0 GiB A502
Command (? for help): q
root@partedmagic:~#
The subsequent start of the computer found the file system and booted the operating system.
On 08/09/16 03:09 AM, Graham Perrin wrote:
I performed another clean installation that was not bootable, then again allowed GParted to fix the GPT.
Before and after the fix – note the change to the last usable sector, from 976773127 to 976773134:
|Welcome - Parted Magic (Linux 3.10.4-pmagic) root@partedmagic:~# gdisk GPT fdisk (gdisk) version 0.8.6 Type device filename, or press
to exit: root@partedmagic:~# gdisk GPT fdisk (gdisk) version 0.8.6 Type device filename, or press to exit: /dev/sda Partition table scan: MBR: protective BSD: not present APM: not present GPT: present Found valid GPT with protective MBR; using GPT. Command (? for help): p Disk /dev/sda: 976773168 sectors, 465.8 GiB Logical sector size: 512 bytes Disk identifier (GUID): 4D7213D0-5DF2-11E6-98FC-68B599FCDB6A Partition table holds up to 128 entries First usable sector is 40, last usable sector is 976773127 Partitions will be aligned on 8-sector boundaries Total free space is 30688 sectors (15.0 MiB) Number Start (sector) End (sector) Size Code Name 1 40 204839 100.0 MiB EF00 2 204840 959965223 457.6 GiB A504 3 959965224 976742439 8.0 GiB A502 Command (? for help): q root@partedmagic:~# date ; uname -a Mon Aug 8 22:56:06 AKDT 2016 Linux partedmagic 3.10.4-pmagic #2 Tue Jul 30 11:06:11 CDT 2013 i686 Intel(R) Core(TM) i5 CPU M 540 @ 2.53GHz GenuineIntel GNU/Linux root@partedmagic:~# gdisk GPT fdisk (gdisk) version 0.8.6 Type device filename, or press to exit: /dev/sda Partition table scan: MBR: protective BSD: not present APM: not present GPT: present Found valid GPT with protective MBR; using GPT. Command (? for help): p Disk /dev/sda: 976773168 sectors, 465.8 GiB Logical sector size: 512 bytes Disk identifier (GUID): 4D7213D0-5DF2-11E6-98FC-68B599FCDB6A Partition table holds up to 128 entries First usable sector is 40, last usable sector is 976773134 Partitions will be aligned on 8-sector boundaries Total free space is 30695 sectors (15.0 MiB) Number Start (sector) End (sector) Size Code Name 1 40 204839 100.0 MiB EF00 2 204840 959965223 457.6 GiB A504 3 959965224 976742439 8.0 GiB A502 Command (? for help): q root@partedmagic:~# | The subsequent start of the computer found the file system and booted the operating system.
— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/trueos/trueos-core/issues/15#issuecomment-238471847, or mute the thread https://github.com/notifications/unsubscribe-auth/AB1_IKOwCRs4pWtwPfaokfVo23qU_WxVks5qeCfHgaJpZM4JehFn.
Ok, since I can't reproduce this here, I'm going to need to rely on you do to some additional debugging. Can you do the following:
Do fresh install of TrueOS. Before you reboot, open terminal and run 'gpart show ada0', send me that output.
Next, do your gparted fix. After you boot TrueOS, send me the same output of 'gpart show ada0'. I need to see what specifically was changed from a gpart perspective, since gparted isn't something we have access to on our install media ;)
Kris Moore iXsystems Enterprise Storage & Servers Driven By Open Source
Fresh, whilst still booted from the USB flash drive:
=> 40 976773088 ada0 GPT (466G)
40 204800 1 efi (100M)
204840 968138752 2 freebsd-zfs (462G)
968343592 8388608 3 freebsd-swap (4.0G)
976732200 40928 - free - (20M)
FreeBSD trueos 12.0-CURRENT FreeBSD 12.0-CURRENT #13 68e1f06(drm-next-4.6): Thu Aug 4 19:31:14 UTC 2016 root@devastator:/usr/obj/usr/src/sys/GENERIC amd64
[bbsadmin-l@momh167-gjp4-hpelitebook8540p-trueos] ~% date ; uptime ; gpart show ada0
Tuesday 9 August 2016 at 19:16:31 BST
7:16p.m. up 9 mins, 0 users, load averages: 1.00, 0.93, 0.55
=> 40 976773095 ada0 GPT (466G)
40 204800 1 efi (100M)
204840 968138752 2 freebsd-zfs (462G)
968343592 8388608 3 freebsd-swap (4.0G)
976732200 40935 - free - (20M)
[bbsadmin-l@momh167-gjp4-hpelitebook8540p-trueos] ~%
This issue was moved to trueos/pc-sysinstall#3
trueos/pc-sysinstall#3 no longer exists :-1: and was referenced by https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=211715
Comparing a recent installer for TrueOS Desktop with the most recent installer for FreeBSD-CURRENT:
Kris, please, can we reopen this issue? And then I'll update the subject line.
I think you might have the same problem I have on a macbook. I posted this on the forum. Check your fdisk layout after installing TrueOS - the EFI/GPT partitions get moved to slice 2. In FreeBSD they stay in slice 1. I found the difference in pc-sysinstall/backend/functions-disk.sh line 715:
rc_halt "gpart set -a lenovofix ${_intDISK}"
I can confirm TrueOS boots now, at least on a macbook. Thanks.
I sense confusion.
Check your fdisk layout after installing TrueOS - the EFI/GPT partitions get moved to slice 2
Triple-checked, this is not the case. With output from fdisk:
$ date ; freebsd-version ; uname -a
Sat 29 Oct 2016 11:21:22 BST
12.0-CURRENT
FreeBSD momh167-gjp4-hpelitebook8540p-trueos.university.brighton.ac.uk 12.0-CURRENT FreeBSD 12.0-CURRENT #12 8b8c812(drm-next-4.7): Wed Oct 5 19:36:49 UTC 2016 root@gauntlet:/usr/obj/usr/src/sys/GENERIC amd64
$
As far as I can tell, the commit above (2016-10-28 15:34 GMT+1) was made after build 109 of the images (2016-10-28 14:35:51) so I'll await the next set of images before confirming whether this bug is fixed.
Today I read the gpart man page again and found an easier fix if you don't want to reinstall. Boot into shell from an installation image (TrueOS or Freebsd) and run:
# gpart unset -a lenovofix ada0
@grahamperrin - did you run "fdisk ada0" from BSD or from gparted? I also didn't see anything wrong in my install under the "other" fdisk :]
fdisk, Linux.
Can I suggest a halt on comments here until after I have tested the next set of images? Thanks.
From multi-issue 114:
… I think I saw on your FreeBSD bug report that if you do a gpart backup / restore it fixes the boot issue? I could throw that into our installer. …
@kmoore134 as a workaround, that's tempting :-) but I reckon, let's wait until after the next set of installers becomes available for testing. I'd like things to become less tangled and work towards a fix that's based upon an agreed understanding of the problem. Thanks.
In the meantime, is it possible for you to access trueos/pc-sysinstall#3 and maybe send to me via e-mail (not yet attach here) a PDF of the content of that issue?
@kmoore134 please advise whether recent https://builds.ixsystems.com/jenkins/view/Builds/view/TrueOS/job/TrueOS%20-%20Build%20ISO/111/ includes the gpart backup/restore routine quoted in my previous comment.
(I don't see anything backup-/restore-related at https://github.com/trueos/pc-sysinstall/commits/master and I don't know whether a related commit might be in a different repo).
Thanks
Aiming to understand discrepancies from what was reported by me:
after installing TrueOS - the EFI/GPT partitions get moved to slice 2
@patryk-s please: was there, in that case, a protective MBR?
I'm sorry but I haven't heard about protective MBR before so I don't know exactly what you mean.
I had this "wont boot after install" problem on two laptops: an old 2006 white Macbook and a Lenovo Yoga 2. The commit that fixed it for me is https://github.com/trueos/pc-sysinstall/commit/dd8ec0446cb319c1a1d4ef31f1a43795d5186784 . I don't have a new installer image obviously, but I just deleted that line from the script before proceeding with the install. I later figured out that you can undo that "lenovofix" after installation with:
gpart unset -a lenovofix ada0
And as I mentioned before, the only way to tell the difference between a working and nonworking boot is with a BSD fdisk. gpart and linux fdisk don't show any differences.
I don't know much about the innards of MBR or UEFI, but that's what worked for me. Anyhow, it would be nice to try out a newer installer image, because the Oct-28 doesn't have the fix.
I reused TrueOS-Desktop-2016-10-28-x64-USB.img for another installation to a MacBook Pro.
Whilst booted from that installer, I'd like to save the output of fdisk to a USB flash drive but (a separate issue) I can't get the OS to mount any suitable drive.
https://www.freebsd.org/cgi/man.cgi?query=fdisk&sektion=8&manpath=FreeBSD+12-current for FreeBSD notes that the fdisk command is obsolete.
https://github.com/trueos/pc-sysinstall/commit/dd8ec0446cb319c1a1d4ef31f1a43795d5186784 fix confirmed with recently published TrueOS-Desktop-2016-12-15-x64-USB.img :+1:
Not tested with the HP EliteBook 8540p, but a fresh installation with defaults is bootable on a MacBookPro8,2.
I am having similar symptoms on a Dell Precision M3800 when trying to reboot with Debian testing... The partition table (GPT+UEFI) disappears somehow.
This happens either: 1) after (successfully) finishing the installation (see https://www.trueos.org/handbook/install.html#installation-finished). 2) (if I restore the partitions afterwards), after each time I boot into TrueOS.
Affected TrueOS versions: 17.12 (stable), 2017.12.30 (unstable). The computer already has Debian testing installed.
BTW: Once I restore the partitions, rEFInd works fine.
This problem only happens after booting into TrueOS. As long as I don't access it, I can use Debian normally.
@kmoore134 Could you please elaborate on how could I open a terminal during install, or on how to log the install process, so I could provide with proper logs, please? I haven't found the relevant info in the TrueOS handbook, and [Alt+{F1-F4}] does not work...
Expected
Load … boot …
Actual result
Then:
Expected: a file system
Actual result: nothing below that heading.
Observations
From recent chats in irc://chat.freenode.net/#pcbsd I understand that tearing is a known issue in some environments. For completeness only, the numbered steps above include steps that are bugged in that way.
The apparent absence of a file system with this EliteBook 8540p is reminiscent of July's bug report for an 8570p and concerning that symptom, Joe Maloney may recall the chat that he and I had in IRC.
BIOS on this 8540p is outdated (68CVD Ver. F.0E of 2010-11-25) but given the same symptom (I guess, the same bug) with the 8570p, I doubt that BIOS is a contributory factor.
References
HP EliteBook 8540p Notebook PC Support - HP Support Center captured 2015-12-08 in the Internet Archive Wayback Machine
HP Notebook System BIOS Update