NetBSD-Bugs archive
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index][Old Index]
kern/56379: raidframe assumes wrong raid components
>Number: 56379
>Category: kern
>Synopsis: raidframe assumes wrong raid components
>Confidential: no
>Severity: critical
>Priority: high
>Responsible: kern-bug-people
>State: open
>Class: sw-bug
>Submitter-Id: net
>Arrival-Date: Sat Aug 28 22:20:00 +0000 2021
>Originator: John Klos
>Release: NetBSD 9.99.88
>Organization:
>Environment:
System: NetBSD athena.zia.io 9.99.88 NetBSD 9.99.88 (GENERIC-$Revision: 1.413 $) #0: Sat Aug 28 15:39:57 UTC 2021 john%frigg.zia.io@localhost:/usr/current/obj-alpha/sys/arch/alpha/compile/HELENA alpha
Architecture: alpha
Machine: alpha
>Description:
After updating to 9.99.88, one of my raidframe mirrors showed duplicate devices:
[ 5.8495935] sd0 at scsibus0 target 0 lun 0: <COMPAQ, BD0726536C, 3B02> disk fixed
[ 5.9413900] sd0: 69464 MB, 28004 cyl, 20 head, 254 sec, 512 bytes/sect x 142264000 sectors
[ 6.0595535] sd0: sync (25.00ns offset 63), 16-bit (80.000MB/s) transfers, tagged queueing
[ 13.5937017] sd1 at scsibus2 target 1 lun 0: <ATA, ST2000LM010-1RA1, SDM1> disk fixed
[ 13.6894044] sd1: 1863 GB, 1907730 cyl, 16 head, 127 sec, 512 bytes/sect x 3907029168 sectors
[ 14.1581536] sd1: tagged queueing
[ 14.2011211] sd2 at scsibus2 target 2 lun 0: <ATA, ST2000LM010-1RA1, SDM1> disk fixed
[ 14.2968242] sd2: 1863 GB, 1907730 cyl, 16 head, 127 sec, 512 bytes/sect x 3907029168 sectors
[ 14.7655725] sd2: tagged queueing
[ 14.8046354] sd3 at scsibus2 target 3 lun 0: <ATA, Samsung SSD 850, 2B6Q> disk fixed
[ 14.8964312] sd3: 465 GB, 476941 cyl, 16 head, 127 sec, 512 bytes/sect x 976773168 sectors
[ 14.9960407] sd3: tagged queueing
[ 15.0351022] sd4 at scsibus2 target 4 lun 0: <ATA, Samsung SSD 850, 2B6Q> disk fixed
[ 15.1268991] sd4: 465 GB, 476941 cyl, 16 head, 127 sec, 512 bytes/sect x 976773168 sectors
[ 15.2265089] sd4: tagged queueing
[ 15.2880306] raid1: RAID Level 1
[ 15.3231865] raid1: Components: /dev/sd1c /dev/sd2c
[ 15.3808034] raid1: Total Sectors: 3907028992 (1907729 MB)
[ 15.7362709] raid0: RAID Level 1
[ 15.7714271] raid0: Components: /dev/sd4a /dev/sd3a
[ 15.8290441] raid0: Total Sectors: 943218560 (460555 MB)
[ 16.2284566] raid2: RAID Level 1
[ 16.2636123] raid2: Components: /dev/sd4c /dev/sd3c
[ 16.3212293] raid2: Total Sectors: 943218560 (460555 MB)
The disklabel for sd3 and sd4 show:
# size offset fstype [fsize bsize cpg/sgs]
a: 943218736 0 RAID # (Cyl. 0 - 935732*)
b: 33554432 943218736 swap # (Cyl. 935732*- 969020)
c: 976773168 0 unused 0 0 # (Cyl. 0 - 969020)
I tried unconfiguring one, then the other, then reconfiguring. Both sets of raid devices for sd3 and sd4 came up every time. I tried reconfiguring, got the same.
I backed up the drives, zeroed the first several dozen gigabytes, wrote a new disklabel:
3 partitions:
# size offset fstype [fsize bsize cpg/sgs]
a: 951605296 2048 RAID # (Cyl. 1*- 468310*)
b: 25165824 951607344 swap # (Cyl. 468310*- 480695*)
c: 976773168 0 unused 0 0 # (Cyl. 0 - 480695*)
Then tried making a new raid mirror specifically using sd3a and sd4a, but I get this:
[ 17.768855] RAIDFRAME: Did not find any live disks in the array.
[ 17.839024] RAIDFRAME: failed rf_ConfigureDisks with 22
[ 17.901430] raid0: raidstart not ready
This is after all proper sets (giving serial, initializing, waiting for parity to finish.
Autoconfigure doesn't work. Manually reconfiguring with raidctl then gives:
raidctl -sv raid0
Components:
/dev/sd3c: optimal
/dev/sd4c: optimal
and
[ 458.109832] raid0: Component /dev/sd3c being configured at col: 0
[ 458.180139] Column: 0 Num Columns: 2
[ 458.232869] Version: 2 Serial Number: 2021082800 Mod Counter: 52
[ 458.313918] Clean: Yes Status: 0
[ 458.361766] raid0: Component /dev/sd4c being configured at col: 1
[ 458.434026] Column: 1 Num Columns: 2
[ 458.485780] Version: 2 Serial Number: 2021082800 Mod Counter: 52
[ 458.567806] Clean: Yes Status: 0
[ 458.618584] raid0: RAID Level 1
[ 458.654714] raid0: Components: /dev/sd3c /dev/sd4c
[ 458.711350] raid0: Total Sectors: 976772992 (476939 MB)
Clearly sd3c and sd4c are wrong, as is the sector count for this raidframe device.
>How-To-Repeat:
>Fix:
>Unformatted:
Home |
Main Index |
Thread Index |
Old Index