[10.0-BETA] Reproductible panic (filesystem)

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

	Hello,

	One of my servers runs NetBSD-10.0 and panics if I try to use a ccd
device after a regular shutdown. I have to manualy run fsck on this
device (automatic fsck aborts), but I'm not sure fsck is enough (I have
seen a panic _after_ fsck). As this device only contains squid cache, I
prefer do a newfs followed by squid -z.

	Server configuration :

legendre:[~] > df -h
Filesystem     Size   Used  Avail %Cap Mounted on
/dev/raid0a     31G    13G    17G  43% /
/dev/raid0e     62G    30G    29G  51% /usr
/dev/raid0f     31G    22G   7,5G  74% /var
/dev/dk5        97G   155M    92G   0% /var/squid/cache
/dev/raid0g    252G   116G   123G  48% /usr/src
/dev/raid0h    523G   300G   197G  60% /srv
/dev/dk0       3,6T   1,5T   2,0T  42% /home
kernfs         1,0K   1,0K     0B 100% /kern
ptyfs          1,0K   1,0K     0B 100% /dev/pts
procfs         4,0K   4,0K     0B 100% /proc
tmpfs          4,0G    48K   4,0G   0% /var/shm
/dev/dk6        11T   9,5T   580G  94% /opt/bacula
/dev/dk7        11T   2,0T   8,0T  20% /opt/video

legendre# dkctl dk5 getwedgeinfo
dk5 at ccd0: squid
dk5: 209715200 blocks at 287309864, type: ffs
legendre# cat /etc/ccd.conf
ccd0    32      none    /dev/wd0a /dev/wd1a

dmesg:
[     1.266369] wd0 at atabus0 drive 0
[     1.266369] wd0: <ST1000LM049-2GH172>
[     1.266369] wd0: drive supports 16-sector PIO transfers, LBA48
addressing
[     1.266369] wd0: 931 GB, 1938021 cyl, 16 head, 63 sec, 512
bytes/sect x 1953525168 sectors (4096 bytes/physsect; first aligned
sector: 8)
...
[     1.716369] wd1 at atabus1 drive 0
[     1.716369] wd1: <ST1000LM049-2GH172>
[     1.716369] wd1: drive supports 16-sector PIO transfers, LBA48
addressing
[     1.716369] wd1: 931 GB, 1938021 cyl, 16 head, 63 sec, 512
bytes/sect x 1953525168 sectors (4096 bytes/physsect; first aligned
sector: 8)
...
[     5.366362] ccd0: Interleaving 2 components (32 block interleave)
[     5.366362] ccd0: /dev/wd0a (1953524160 blocks)
[     5.366362] ccd0: /dev/wd1a (1953524160 blocks)
[     5.366362] ccd0: total 3907048320 blocks
[     5.366362] ccd0: GPT GUID: 93510c33-2f8d-45d4-8385-21d894397de6
[     5.366362] dk1 at ccd0: "swap_hilbert", 201326592 blocks at 40,
type: swap
[     5.366362] dk2 at ccd0: "swap_abel", 2097152 blocks at 201326632,
type: swap
[     5.366362] dk3 at ccd0: "swap_schwarz", 33554432 blocks at
203423784, type: swap
[     5.366362] dk4 at ccd0: "swap_pythagore", 50331648 blocks at
236978216, type: swap
[     5.366362] dk5 at ccd0: "squid", 209715200 blocks at 287309864,
type: ffs
...

	Both wd0 and wd1 have been tested and don't have smart alerts. When I
restart this server, I can see a message just before system reboots. If
I remember, this message indicates that ccd0 is removed from system
before dk5 and could explain why filesystem is corrupted after next boot.

	Same ccd0 device ran fine with -9.3 (but iSCSI initiator was broken, I
use -10.0 as a lot of iSCSI bugs have been fixed. Thanks.).

	Best regards,

	JKB