Resending due to "Mailinglist outage yesterday" - was sent at 12:42 UTC 18/7
On 18/07/2019 08.40, stakanov wrote:
In data giovedì 18 luglio 2019 01:42:08 CEST, Carlos
E. R. ha scritto:
On 17/07/2019 09.00, stakanov wrote:
In data mercoledì 17 luglio 2019 08:38:46 CEST,
Felix Miata ha scritto:
stakanov composed on 2019-07-17 08:18
> more or less that is what I find.
> the machine was running for long time and was suspended before the
> NVRAM may be full? How could I free it in order to boot?
Surplus entries may be removed using the efibootmgr utility in a rescue
mode boot. Simply enter efibootmgr in a root login shell to list all the
entries. Man efibootmgr shows the options required to remove unwanted
Thank you. I checked also in the bios.
I have to say that it is very(!) misleading having this error in the log
when you have (like in this case) another issue.
In fact I had a ext4 problem on /home that would not mount but the error
shown first in the logs and in the boot process is that stupid could not
get uefi list.
You said the machine was suspended?
Suspended means suspended to RAM, different than hibernation, where the
image is saved to disk. If the battery runs out while suspended, on next
boot the machine has to recover from bad corruption on all the mounted
partitions at the time of suspend.
I did not know that. Yes, it was "suspend
to ram" and yes, it may have been
that the suspended partition was not correctly recovered when the update
required a reboot. So the corruption may be a result of this.
Thanks for the heads up.
I'm not sure if you understood the issue, so I'll try to explain again,
just in case - maybe it is a translation issue :-)
The update is not related.
The machine gets suspended to ram; all the mounted partitions are, well,
mounted: opened files, half written files, the lot. The battery doesn't
last enough, and the machine, instead of recovering from suspend, reboots.
The effect is basically the same as recovering from a hard crash in the
middle of the session. Every partition in use is corrupt. Some may
recover fine, some not.
There is a suspend method that avoids this, called "hybrid". It does
both suspend to ram and to disk procedures. On recover, if battery
lasts, it just continues from ram, instant recover. If the battery
failed, it recovers from disk just as from a normal hibernation.
Whether the machine was updated doesn't affect any of this. So if you
suspended, recovered, updated, and rebooted, then that was just normal work.
However, I still get the "could not get uefi list" while the uefi entries are
correct and no residuals are to be found. I think that that error is because
the machine has secure boot without a TPM.
Trusted Platform Module?
If a command can read the uefi list, I don't see why some part of the
boot process can't.
Cheers / Saludos,
Carlos E. R.
(from 15.0 x86_64 at Telcontar)