the effect of this is cosmetic, in that we cannot obtain the space
allocated to each LUN. A user on the new version (0.6.2) of
drobo-utils using it on the PRO. He reports the following output:
for 'drobom info luns'
lun size (GB) used PTFmt FStype:
0 2199 2000 GPT ['EXT3']
1 2199 0 GPT ['EXT3']
2 2199 0 GPT ['EXT3']
2000 is the total used on all three LUNS, rather than the first one.
The space is distributed among the LUNS. so the third column is inaccurate.
The third column is the response supplied by the Drobo to a
Drobo LUNINFO (not LUNINFO2) mode sense query for each LUN.
section 3.8.1 for the Drobo Management interface Protocol (DMIP) document
indicates this field is:
"The LUN Used Capacity is the used capacity of this particular LUN."
I reproduced the behaviour on a v1 with two luns, sg1 (the second 1 TB LUN)
has 7G's used, versus the first LUN being virtually empty. Yet all usage is
reported against the first LUN, not the second one.
root@pepino:~# df /mnt2 /mnt
Filesystem 1K-blocks Used Available Use% Mounted on
/dev/sdf1 1072622668 131224 1072491444 1% /mnt2
/dev/sdg1 1072622668 7411432 1065211236 1% /mnt
Post: #1 LUNINFO modesense response wrong?
Hi folks,
the effect of this is cosmetic, in that we cannot obtain the space
allocated to each LUN. A user on the new version (0.6.2) of
drobo-utils using it on the PRO. He reports the following output:
for 'drobom info luns'
lun size (GB) used PTFmt FStype:
0 2199 2000 GPT ['EXT3']
1 2199 0 GPT ['EXT3']
2 2199 0 GPT ['EXT3']
2000 is the total used on all three LUNS, rather than the first one.
The space is distributed among the LUNS. so the third column is inaccurate.
The third column is the response supplied by the Drobo to a
Drobo LUNINFO (not LUNINFO2) mode sense query for each LUN.
section 3.8.1 for the Drobo Management interface Protocol (DMIP) document
indicates this field is:
"The LUN Used Capacity is the used capacity of this particular LUN."
I reproduced the behaviour on a v1 with two luns, sg1 (the second 1 TB LUN)
has 7G's used, versus the first LUN being virtually empty. Yet all usage is
reported against the first LUN, not the second one.
root@pepino:~# df /mnt2 /mnt
Filesystem 1K-blocks Used Available Use% Mounted on
/dev/sdf1 1072622668 131224 1072491444 1% /mnt2
/dev/sdg1 1072622668 7411432 1065211236 1% /mnt
root@pepino:~# drobom info
---------------------------------------------------------
Drobo Name: hoho Devices: /dev/sdf:/dev/sdg
Time: Mon Dec 28 10:57:08 2009
---------------------------------------------------------
Configuration maxima: slots: 4, luns: 16, lunsize: 1 TB
Capacity (in GB): used: 10, free: 1484, total: 1494
protocol version: 0.11
SCSI emulation information:
Vendor: TRUSTED Model: Mass Storage Revision: 1.00
/dev/sdf: bus=scsi18 channel=0 id=0 lun=0 (vendor=TRUSTED , version=30527)
/dev/sdg: bus=scsi18 channel=0 id=0 lun=1 (vendor=TRUSTED , version=30527)
query slotinfo result: number of slots: 4
slot GB Model Status
0 500 ST3500830AS green
1 750 WDC WD7500AAKS-00RBA0 green
2 500 WDC WD5000AAKS-00C8A0 green
3 500 WDC WD5000AAKS-00C8A0 green
Firmware: 1.3.5
Revision: 1.252 ( 21110 ) built: Aug 31 2009,18:12:03
Features: NO_AUTO_REBOOT,NO_FAT32_FORMAT,USED_CAPACITY_FROM_HOST,DISKPACKSTATUS,ENCRYPT_NOHEADER,CMD_STATUS_QUERIABLE,VARIABLE_LUN_SIZE_1_16,PARTITION_LUN_GPT_MBR,FAT32_FORMAT_VOLNAME,SUPPORTS_DROBOSHARE,SUPPORTS_NEW_LUNINFO2,feature x0800,feature x2000
/dev/sdf:/dev/sdg /mnt:/mnt2 hoho 00% full - ([], 0)
query options result:
RedThreshold 95
YellowThreshold 85
lun size (GB) used PTFmt FStype:
0 1099 10 GPT ['EXT3']
1 1099 0 GPT ['EXT3']
---------------------------------------------------------
root@pepino:~#
It appears that Drobo is not reporting the amount of space assigned to each LUN,
but reporting all space as belonging to the first LUN.
This is caused by a bug in the underlying DRI firmware.
reported to DRI via drobospace...
http://www.drobospace.com/forums/showthread.php?tid=772