| Summary: | If EFI partition is mounted during custom partitioning draklive-install fails with "grub2-install: error: cannot find EFI directory." | ||
|---|---|---|---|
| Product: | Mageia | Reporter: | Dick Gevers <dvgevers> |
| Component: | RPM Packages | Assignee: | Mageia tools maintainers <mageiatools> |
| Status: | RESOLVED FIXED | QA Contact: | |
| Severity: | normal | ||
| Priority: | Normal | CC: | isobuild, mageia, mageiatools, marja11, stormi-mageia, zen25000 |
| Version: | 6 | Keywords: | 6.1, FOR_ERRATA6 |
| Target Milestone: | --- | ||
| Hardware: | x86_64 | ||
| OS: | Linux | ||
| Whiteboard: | |||
| Source RPM: | drakxtools? | CVE: | |
| Status comment: | Seen with Mageia-6.1-rc-LiveDVD-GNOME-x86_64-DVD.iso | ||
| Attachments: | Result of journalctl -a (and xz) ... after crash of draklive-install | ||
|
Description
Dick Gevers
2018-08-06 12:21:02 CEST
@ stormi or LpSolit Can one of you please add the "6.1RC" keyword for the isotesters and isobuilders? @ Dick Assuming the crash didn't cause a hard freeze: Can you please reproduce the issue and fetch the logs? As root: journalctl -a > log.txt compress if needed: xz log.txt and attach log.txt.xz (or log.txt) to this report Source RPM:
LiveDVD Gnome 64 bits =>
draklive-install? drakxtools? grub2? (In reply to Marja Van Waes from comment #2) > Assuming the crash didn't cause a hard freeze: Can you please reproduce the > issue and fetch the logs? Sure, but not before 1600 h today ;) I've created the 6.1 keyword (we used a similar 5.1 keyword in the past, without the need for 5.1RC), that can be used for any bug specific to the 6.1 ISOs. Keywords:
(none) =>
6.1
Frédéric "LpSolit" Buclin
2018-08-07 14:06:52 CEST
CC:
LpSolit =>
(none) Reproduced today (yesterday testroom was too hot ;(( ) "drakbug: showed follg on screen: quote The "draklive-install" program has crashed with the following error: grub2-install failed: Installing for x86_64-efi platform. grub2-install: error: cannot find EFI directory. ...propagated at /usr/lib/libDrakX/any.pm line 269. Perl's trace: drakbug::bug_handler() called from /usr/lib/libDrakX/any.pm:269 any::installBootloader() called from /usr/lib/libDrakX/any.pm:239 any::setupBootloaderUntilInstalled() called from /usr/sbin/draklive-install:347 main::setup_bootloader() called from /usr/sbin/draklive-install:72 main::install_live() called from /usr/sbin/draklive-install:44 Used theme: Adwaita unquote Will attach log as per Marja's instructions Created attachment 10305 [details]
Result of journalctl -a (and xz) ... after crash of draklive-install
When you mounted it, there was this message: Aug 08 08:54:58 dvglt2.fritz.box kernel: FAT-fs (sda9): Volume was not properly unmounted. Some data may be corrupt. Please run fsck. But, if some data had gotten corrupted, shouldn't the next draklive-install run, without manually mounting the ESP, then have failed, too? Anyway, does fsck find anything to repair if you run it on /dev/sda9 ? Summary:
If EFI partition is mounted during custom partitioning draklive-install fails with "cannot find EFI directory" =>
If EFI partition is mounted during custom partitioning draklive-install fails with "grub2-install: error: cannot find EFI directory." (In reply to Marja Van Waes from comment #7) > When you mounted it, there was this message: > > Aug 08 08:54:58 dvglt2.fritz.box kernel: FAT-fs (sda9): Volume was not > properly unmounted. Some data may be corrupt. Please run fsck. > > But, if some data had gotten corrupted, shouldn't the next draklive-install > run, without manually mounting the ESP, then have failed, too? Indeed 2 subsequent draklive-installs succeeded (without manual run of fsck). > Anyway, does fsck find anything to repair if you run it on /dev/sda9 ? Yes, at this time running fsck.vfat it shows ...0x41 Dirty bit set. Fs was not properly unmounted and some data may be corrupt... (In reply to Dick Gevers from comment #8) > Yes, at this time running fsck.vfat it shows > ...0x41 Dirty bit set. Fs was not properly unmounted and some data may be > corrupt... I see the same on my main system, also with my FAT partition. The few times I ever boot Windows, it complains and fixes it, but it always comes back. I've just tried repairing the fault with fsck.vfat, but it's back after a reboot. I've never worried about this because I don't keep anything valuable on my FAT partition, and anyway the FAT filesystem generally survives unless you are actually writing files at the time. We should have a separate bug report for this if you think it's worth pursuing. (In reply to Martin Whitaker from comment #9) > We should have a separate bug report for this if you think it's worth > pursuing. Sorry, but given that i stopped having any windows system 15 years ago i don't think i could be much help. The only windowsy thing on 5 machines is this EFI (vfat) partition I needed on this laptop because the Mageia installer wouldn't proceed without one when I bought the machine last fall.... I did write/remove some data from the EFI partition and each time unmounted/mounted it with diskdrake to see if this would corrupt the filesystem by any chance, but running "fsck.vfat -b -n -v -V..." afterwards does not show corruption now, so the diskdrake function in the installer is presumably okay (this being diskdrake from drakxtools...17.98-2.mga7) If I keep the EFI partition not mounted (unmounted through diskdrake) and run grub-customizer from Cauldron installed on the testmachine I get a very similar error when I try to save the bootloader to MBR: quote Error while installing the bootloader Installing for x86_64-efi platform. grub2-install: error: cannot find EFI directory. [ OK ] unquote so maybe the error is caused by either grub2-efi or by the way that draklive-install passes commands to grub2(-efi).
Marja Van Waes
2018-08-10 18:06:53 CEST
CC:
(none) =>
zen25000
Dick Gevers
2018-08-16 12:52:24 CEST
Keywords:
(none) =>
FOR_ERRATA6 The fix for bug 23603 fixes this bug too. This should be fixed on the ISOs dated 1st October. Yes looks fine to me Status:
NEW =>
RESOLVED |