.sujournal strangeness after (?) upgrading to 10-STABLE

[Jakub Lach]

Unfortunately, I had my first reboot. Not trusting journal (I've enabled it
on all filesystems some time ago) I did full few fscks between reboots 
in single user mode.

Unsurprisingly, there were some inconsistencies in / and /var _but_
checking tunefs -p, actually SU+J was disabled there (on all other
filesystems it was still enabled)!

I don't know how that happened, previously I just didn't trust SU+J
enough and did fsck anyway, I hope that explains inconsistencies.

For some time I couldn't enable SU+J again, because of  .sujournal
actaully present both in / and /var - so there was evidence I really 
enabled it but somehow it disabled itself? 

Anyway, using chflags 0 I've removed both journals and after fsck was
successful in enabling both journals again by tunefs -j enable 
(/ after a reboot with "read only fs modified" error).

However, that new journals (initial ones were created somewhere on 
9-STABLE I think) bear an actual date of enabling them- in difference 
to the old ones, that have just a start of Unix epoch (1970). 

Is it intended behaviour?

/tmp $ ls -lioF .sujournal                                        
3 -r--------  1 root  wheel  schg,sunlnk,nodump,opaque 33554432  1 sty  1970
.su
/tmp $ cd /var
/var $ ls -lioF .sujournal                                        
4 -r--------  1 root  wheel  schg,sunlnk,nodump,opaque 33554432 28 sty 02:06
.su




--
View this message in context: http://freebsd.1045724.n5.nabble.com/sujournal-strangeness-after-upgrading-to-10-STABLE-tp5880684.html
Sent from the freebsd-stable mailing list archive at Nabble.com.