Switch to tty1 multi-times in ro snapshot #5629
Conversation
mitiao
force-pushed
the
poo#40154
branch
2 times, most recently
from
6036501
to
a0b98ab
Compare
lib/migration.pm
Outdated
| else { | ||
| record_info('not in tty1', 'switch to tty1 failed', result => 'softfail'); | ||
| } | ||
| die "Boot into read-only snapshot failed over 5 minutes, considering a product issue" if $i == $retries; |
There was a problem hiding this comment.
You could simply put this out of the loop unconditionally to avoid check on every iteration and return once tty1-selected matches.
There was a problem hiding this comment.
yeah, i have optimized a little
lib/migration.pm
Outdated
| record_info('not in tty1', 'switch to tty1 failed', result => 'softfail'); | ||
| } | ||
| die "Boot into read-only snapshot failed over 5 minutes, considering a product issue" if $i == $retries; | ||
| sleep 10; |
There was a problem hiding this comment.
You could also put this in checkscreen, and have timeout there, so we have chance to match it earlier.
There was a problem hiding this comment.
use checkscreen here is no sense, tty switch by sending key.
There was a problem hiding this comment.
Sorry for the confusion, that's not what I mean, there is already check-screen above, instead of sleep we could wait longer to match tty1 instead of idle time.
Imagine following, tty1 appear in 5 seconds, with your current implementation we will wait 12 seconds before we match it, if we put 10 seconds, we will match it immediately.
In case it doesn't appear we will still wait 10 seconds before retry. Hope the point is clear now.
There was a problem hiding this comment.
yes, hate checkscreen :)
This may fix some sporadic failures during booting read-only snapshot at snapper rollback stage. See: https://progress.opensuse.org/issues/40154
This MAY fix some sporadic failures during booting read-only snapshot at snapper rollback stage.
The failure is hard to reproduce, sometimes it would not appear if we give another run.