Re: [10.0-BETA] Reproductible panic (filesystem)

[BERTRAND Joël <joel.bertrand%systella.fr@localhost>]

BERTRAND Joël a écrit :
> BERTRAND Joël a écrit :
>> Taylor R Campbell a écrit :
>>>> Date: Fri, 19 May 2023 11:55:09 +0200
>>>> From: BERTRAND Joël <joel.bertrand%systella.fr@localhost>
>>>>
>>>> 	This morning, ccd0 (and of course dk5) are "locked". All tasks stall in
>>>> uninterruptible state when they try to access to dk5:
>>>>
>>>> legendre# cd /var/squid/cache
>>>> legendre# du -hs
>>>> ^C^C^C^C^Cs^C^C^C
>>>
>>> Can you show the output of ^T?  Can you enter crash(8) or ddb and get
>>> process listings and stack traces?
>>
>> 	No, sorry, I have restarted whole system but all process that try to
>> access to dk5 (ffs2 on a ccd disk slice) remain in D state. I will try
>> next time.
> 
> 	Hello,
> 
> 	I have tried last -10.0 BETA and this issue don't seem to be fixed. ccd
> seems to be unmonted _before_ filesystems. On next reboot, a manuel fsck
> complains and refuse to clean filesystem.
> 
> 	I've never seen other panics since system doesn't mount dk5 (I always
> try to run manually fsck).
> 
> 	Best regards,
> 
> 	JKB
> 

	Hello,

	With last -10.0-BETA (I haven't done fsck or newfs on last reboot) :

[ 665263,240089] panic: /var/squid/cache: bad dir ino 459520 at offset
34157: null entry

[ 665263,240089] cpu3: Begin traceback...
[ 665263,240089] vpanic() at netbsd:vpanic+0x183
[ 665263,240089] panic() at netbsd:panic+0x3c
[ 665263,240089] ufs_lookup() at netbsd:ufs_lookup+0x442
[ 665263,240089] VOP_LOOKUP() at netbsd:VOP_LOOKUP+0x3d
[ 665263,240089] lookup_once() at netbsd:lookup_once+0x193
[ 665263,240089] namei_tryemulroot.constprop.0() at
netbsd:namei_tryemulroot.constprop.0+0x373
[ 665263,240089] namei() at netbsd:namei+0x43
[ 665263,240089] vn_open() at netbsd:vn_open+0xff
[ 665263,250087] do_open() at netbsd:do_open+0xc3
[ 665263,250087] do_sys_openat() at netbsd:do_sys_openat+0x74
[ 665263,250087] sys_open() at netbsd:sys_open+0x24
[ 665263,250087] syscall() at netbsd:syscall+0x1fc
[ 665263,250087] --- syscall (number 5) ---
[ 665263,250087] netbsd:syscall+0x1fc:
[ 665263,250087] cpu3: End traceback...


legendre# fsck /dev/dk5
** /dev/rdk5
** File system is journaled; replaying journal
** Last Mounted on /var/squid/cache
** Phase 1 - Check Blocks and Sizes
7912408300994173476 BAD I=3
4313599915630302063 BAD I=3
-4473632163892877928 BAD I=3
8068741989830080453 BAD I=3
3857159125896022134 BAD I=3
-4354179704011695453 BAD I=3
7611175298055105740 BAD I=3
3985638883347136889 BAD I=3
-2495754894521232470 BAD I=3
7739654885841380823 BAD I=3
3564105388769654272 BAD I=3
EXCESSIVE BAD BLKS I=3
CONTINUE? [yn]

INCORRECT BLOCK COUNT I=3 (131168 should be 49888)
CORRECT? [yn] y

INCORRECT BLOCK COUNT I=459520 (8 should be 32)
CORRECT? [yn] y

1895866 DUP I=459689
1895867 DUP I=459690
1895868 DUP I=459691
1895869 DUP I=459692
1895870 DUP I=459693
1895871 DUP I=459694
UNKNOWN FILE TYPE I=528560
...

	Of course, hardware has no issue and system runs fine until it is rebooted.

	Best regards,

	JKB