[dm-crypt] Cryptsetup bug

Sven Eschenberg sven at whgl.uni-frankfurt.de
Fri Dec 30 12:45:19 CET 2016

Since sdb1 has a UUID that differes from the one that was just created 
it seems to be a stale one.

If you want to verify what's going on and you are sure you can reproduce 
it by the same steps then do the following:

Create your partitions, then verify there's no stale signatures.
First verify the kernel reflects your partitioning on the disk.
Then run blkid.
If there's signatures, use something like wipefs to remove them (or dd 
Then recreate the container on sdb2 by issuing your appropriate command.
And run blkid again, to verify only on sdb2 a signature was created.

If suddendly a new luks signature would pop up on sdb1, you can start 
stracing the problem.



Am 30.12.2016 um 03:46 schrieb Hunter DeMeyer:
> /dev/sdb1: UUID="fc1abc36-a8f8-4ace-b324-77f9ddd037ea"
> TYPE="crypto_LUKS" PARTLABEL="EFI System"
> PARTUUID="bacaef0b-61a8-4460-98e7-8a2251d78268"
> /dev/sdb2: UUID="8f6b6ff5-2f49-4edf-a35b-0e8df82e1e44"
> TYPE="crypto_LUKS" PARTLABEL="Linux filesystem"
> PARTUUID="a58cce87-f792-4e5b-8c35-b4210472f86c"
> On Thu, Dec 29, 2016 at 9:18 PM, Milan Broz <gmazyland at gmail.com
> <mailto:gmazyland at gmail.com>> wrote:
>     On 12/29/2016 06:53 PM, Hunter DeMeyer wrote:
>     > I believe that under certain circumstances cryptsetup will encrypt
>     > more than one partion on a drive. I have attached the process that
>     > produced this bug.
>     Cryptsetup accesses only device that you specify.
>     There are two options how could happen that LUKS device appeared on
>     different partition:
>     1) there was some old LUKS device on the /dev/sdb1 offset start
>     already or
>     2) you recreated partitions but for some reason kernel still see the
>     old partitions size,
>     and sdb1 was still in fact mapped to the old partition.
>     Could you please paste blkid /dev/sdb* to verify it is the first
>     option (to check LUKS UUID)?
>     I guess you will see different UUIDs.
>     And please do not send html mails to the list.
>     Milan
> _______________________________________________
> dm-crypt mailing list
> dm-crypt at saout.de
> http://www.saout.de/mailman/listinfo/dm-crypt

More information about the dm-crypt mailing list