gvinum looses drives

From: Matthias Schuendehuette <msch_at_snafu.de>
Date: Sat, 19 Jun 2004 12:47:43 +0200
Hi Lukas et al,

I decided to be brave and give geom_vinum a try...

Here the typescript of the test with my comments inserted:

------------------------8><-------------------------------
Script started on Sat Jun 19 12:25:24 2004

root_at_current - ~
501 # gvinum
gvinum -> list
5 drives:
D testdrive             State: up	/dev/da4	A: 15/2063 MB (0%)
D d1                    State: up	/dev/da0s3	A: 8/4133 MB (0%)
D d2                    State: up	/dev/da1s1	A: 227/4353 MB (5%)
D d3                    State: up	/dev/da2s1	A: 231/4356 MB (5%)
D d4                    State: up	/dev/da3s2	A: 478/4604 MB (10%)

2 volumes:
V mp3dev                State: up	Plexes:       1	Size:       2048 MB
V raid5                 State: up	Plexes:       1	Size:         12 GB

2 plexes:
P mp3dev.p0           C State: up	Subdisks:     1	Size:       2048 MB
P raid5.p0           R5 State: up	Subdisks:     4	Size:         12 GB

5 subdisks:
S mp3dev.p0.s0          State: up	D: testdrive    Size:       2048 MB
S raid5.p0.s3           State: up	D: d4           Size:       4125 MB
S raid5.p0.s2           State: up	D: d3           Size:       4125 MB
S raid5.p0.s1           State: up	D: d2           Size:       4125 MB
S raid5.p0.s0           State: up	D: d1           Size:       4125 MB
gvinum -> quit

# Well, looks good! All items found und "up 'n runnin' "
# Now mount it!

root_at_current - ~
502 # mount /dev/gvinum/mp3dev /mp3dev
mount: /dev/gvinum/mp3dev: Device not configured

# Oops... what's that?

root_at_current - ~
503 # gvinum list
5 drives:
D testdrive             State: up	/dev/da4	A: 15/2063 MB (0%)
D d1                    State: up	/dev/da0s3	A: 8/4133 MB (0%)
D d2                    State: up	/dev/da1s1	A: 227/4353 MB (5%)
D d3                    State: up	/dev/da2s1	A: 231/4356 MB (5%)
D d4                    State: up	/dev/da3s2	A: 478/4604 MB (10%)

2 volumes:
V mp3dev                State: down	Plexes:       1	Size:       2048 MB
V raid5                 State: up	Plexes:       1	Size:         12 GB

2 plexes:
P mp3dev.p0           C State: down	Subdisks:     1	Size:       2048 MB
P raid5.p0           R5 State: up	Subdisks:     4	Size:         12 GB

5 subdisks:
S mp3dev.p0.s0          State: stale	D: testdrive    Size:       2048 MB
S raid5.p0.s3           State: up	D: d4           Size:       4125 MB
S raid5.p0.s2           State: up	D: d3           Size:       4125 MB
S raid5.p0.s1           State: up	D: d2           Size:       4125 MB
S raid5.p0.s0           State: up	D: d1           Size:       4125 MB

# Hmm... 'testdrive' is 'up' but 'mp3dev.p0.s0' is 'stale'. As you see 
# below, geom_vinum lost 'testdrive' but tells still that it's 'up'...

root_at_current - ~
504 # mount /dev/gvinum/raid5 /raid/ 
mount: /dev/gvinum/raid5: Device not configured

# Same here...

root_at_current - ~
505 # gvinum list
5 drives:
D testdrive             State: up	/dev/da4	A: 15/2063 MB (0%)
D d1                    State: up	/dev/da0s3	A: 8/4133 MB (0%)
D d2                    State: up	/dev/da1s1	A: 227/4353 MB (5%)
D d3                    State: up	/dev/da2s1	A: 231/4356 MB (5%)
D d4                    State: up	/dev/da3s2	A: 478/4604 MB (10%)

2 volumes:
V mp3dev                State: down	Plexes:       1	Size:       2048 MB
V raid5                 State: down	Plexes:       1	Size:         12 GB

2 plexes:
P mp3dev.p0           C State: down	Subdisks:     1	Size:       2048 MB
P raid5.p0           R5 State: down	Subdisks:     4	Size:         12 GB

5 subdisks:
S mp3dev.p0.s0          State: stale	D: testdrive    Size:       2048 MB
S raid5.p0.s3           State: stale	D: d4           Size:       4125 MB
S raid5.p0.s2           State: stale	D: d3           Size:       4125 MB
S raid5.p0.s1           State: stale	D: d2           Size:       4125 MB
S raid5.p0.s0           State: stale	D: d1           Size:       4125 MB

# The drives are 'up' but the subdisks are 'stale'

root_at_current - ~
507 # tail -n 256 /var/log/messages
Jun 19 12:24:54 current kernel: vinum: unloaded
Jun 19 12:25:39 current kernel: FOO: sd raid5.p0.s3 is up
Jun 19 12:25:39 current kernel: FOO: sd raid5.p0.s2 is up
Jun 19 12:25:40 current kernel: FOO: sd raid5.p0.s1 is up
Jun 19 12:25:40 current kernel: FOO: sd raid5.p0.s0 is up
Jun 19 12:25:40 current kernel: FOO: sd mp3dev.p0.s0 is up
Jun 19 12:26:47 current kernel: gvinum: lost drive 'testdrive'
Jun 19 12:26:47 current kernel: FOO: sd mp3dev.p0.s0 is down
Jun 19 12:26:47 current kernel: FOO: plex mp3dev.p0 is down
Jun 19 12:26:47 current kernel: FOO: sd mp3dev.p0.s0 is stale
Jun 19 12:26:47 current kernel: FOO: plex mp3dev.p0 is down
Jun 19 12:27:28 current kernel: gvinum: lost drive 'd4'
Jun 19 12:27:28 current kernel: FOO: sd raid5.p0.s3 is down
Jun 19 12:27:28 current kernel: FOO: plex raid5.p0 is degraded
Jun 19 12:27:28 current kernel: gvinum: lost drive 'd3'
Jun 19 12:27:28 current kernel: FOO: sd raid5.p0.s2 is down
Jun 19 12:27:28 current kernel: FOO: plex raid5.p0 is down
Jun 19 12:27:28 current kernel: gvinum: lost drive 'd2'
Jun 19 12:27:28 current kernel: FOO: sd raid5.p0.s1 is down
Jun 19 12:27:28 current kernel: FOO: plex raid5.p0 is down
Jun 19 12:27:28 current kernel: gvinum: lost drive 'd1'
Jun 19 12:27:28 current kernel: FOO: sd raid5.p0.s0 is down
Jun 19 12:27:28 current kernel: FOO: plex raid5.p0 is down
Jun 19 12:27:28 current kernel: FOO: sd raid5.p0.s3 is stale
Jun 19 12:27:28 current kernel: FOO: plex raid5.p0 is down
Jun 19 12:27:28 current kernel: FOO: sd raid5.p0.s2 is stale
Jun 19 12:27:28 current kernel: FOO: plex raid5.p0 is down
Jun 19 12:27:28 current kernel: FOO: sd raid5.p0.s1 is stale
Jun 19 12:27:28 current kernel: FOO: plex raid5.p0 is down
Jun 19 12:27:28 current kernel: FOO: sd raid5.p0.s0 is stale
Jun 19 12:27:28 current kernel: FOO: plex raid5.p0 is down

root_at_current - ~
508 # exit

Script done on Sat Jun 19 12:28:19 2004
------------------------8><-------------------------------

So, I'm quite sure you need additional informations - please tell me 
what.

BTW: Your fix of the unload-preventing geom-bug: will it be commited?
-- 
Ciao/BSD - Matthias

Matthias Schuendehuette	<msch [at] snafu.de>, Berlin (Germany)
PGP-Key at <pgp.mit.edu> and <wwwkeys.de.pgp.net> ID: 0xDDFB0A5F
Received on Sat Jun 19 2004 - 08:48:06 UTC

This archive was generated by hypermail 2.4.0 : Wed May 19 2021 - 11:37:58 UTC