Since 2.12r-14 hwclock runs intentionally after checkroot to be able to
write adjtime. But checkroot "Should-Start"-s hwclockfirst (Petter, note
that this should be changed to hwclock when you have a chance). So
there's an issue.
The current order is causing [worrying] failed fscks.
Basically, on many systems (it was 2 boxes on 2 tested here, probably
all boxes with local RTC and in a timezone earlier than Greenwich), fsck
often takes twice to run ; once to "FIX" the last mount time, and one to
record a successful check. The second, successful fsck is reassuring.
Nevertheless, if you have an attentive admin watching the boot sequence,
that happens to be watching in the five seconds between the failure and
the reboot and which has the time to read that fsck failed, he may be
quite worried about what happened, and not be completely reassured by
the second, successful fsck. And if he's like me, he may spend a few
hours debugging it.
Appart from that, it causes wrong "Last mount/write time" after the "FIX".
I would not dislike seeing that fixed for Etch.
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
