Re: Failed to start mysql due to start_lsn and ens_lsn issue

[Jakub Lopuszanski]

Assuming that you have a backup of the data-dir, and therefore can experiment by starting the mysqld and if something goes wrong restore the data-dir to original state, I'd recommend you try to remove just the #ib_redo27250 and start again in the usual way (without any innodb_force_recovery option!).
The rationale for this particular experiment is, that apparently the system crashed in a state which *looks as if* #ib_redo27249 was not yet full, so it might boot, provided there's no #ib_redo27250 file.
However, the sad news is that we also see clearly that before the crash the system already tried to create the #ib_redo27250 file, which means it was convinced that #ib_redo27249 was full - this means there are some changes to both #ib_redo27249 and #ib_redo27250 that got lost :/
The reason why they've got lost is still unclear to me - I see that at least in 2009 there were issues with LVM not handling fsync() well, but in 2020 they should've been fixed already.
Still, assuming that this "suffix of #ib_redo27249 and prefix of #ib_redo27250" are already lost, the next best thing you can try to do is try to boot without #ib_redo27250 file.