Creating raidz-volume shows low capacity with SSD-Cache

[propeller]

Hi,
- Supermicro board, Supermicro Case
- 2 LSI 9211-8i Controllers
- 12 WD SATA 2 Terabyte Disks
- 2 Samsung S830 128 GByte SATA 6 GByte (for Cache)
- setting up from web-Gui several raidz or raidz2 volumes -> works well as long as only the WD 2 TByte Disks are involved

Problems
- only one SSD is shown in the web-gui, don't know command line to show more, tried severel setting in the LSI Bios,
- tried to connect one to the other controller, same result, in the firmware-Bios all disks are shown
- the LSI-firmware is updated to v11 IR, did not update to IT-Firmware, because I read here in this forum elsewhere, perhaps the SAS supermicro - backplane works not correct (not sure) with IT
- including the one SSD in a raidz-volume (I think zpool ist the correct name) shows a healthy volume but only a few Gigabyte capacity (eg. 476 GByte vs. 7,6 TByte (not the exact values, only relation))
- tried this with mark as "cache" (what is intended) or with "none" in GUI whith same result
- tried the 2 dd if=/dev/zero ... -lines found in this forum for each drive to "clear" them, but no other result.
- tried to create volumes or datasets on the above zpool with reservation '1t' gives me an error "out of space"

- editing disk-properties shows "auto" for the transfer-modus up to 3 Gbit/s (SATA 2) but no 6 Gbit/s, which should work for the samsum 830 SSDs



- tried not another SSD (which should be next try)
- any other ideas?

thank you
uwe

 

[William Grzybowski]

post output of zpool status

 

[propeller]

"cache" command n

Thank you for reply.
The version is 8.02
my (zpool) status is healthy ;)

after destroying and create it with gui (marked da0 as "cache") it looks so:

Code:

[root@fnas1] ~# zpool list
NAME   SIZE   USED  AVAIL    CAP  HEALTH  ALTROOT
zp1    476G   138K   476G     0%  ONLINE  -

on command line it looks so:

Code:

[root@fnas1] ~# zpool create zp1 raidz da0 da1 da2 da3
invalid vdev specification
use '-f' to override the following errors:
raidz contains devices of different sizes

once again

Code:

[root@fnas1] ~# zpool create -f zp1 raidz da0 da1 da2 da3


[root@fnas1] ~# zpool list
NAME   SIZE   USED  AVAIL    CAP  HEALTH  ALTROOT
zp1    476G   138K   476G     0%  ONLINE  -

[root@fnas1] /# zpool status
  pool: zp1
 state: ONLINE
 scrub: none requested
config:

        NAME        STATE     READ WRITE CKSUM
        zp1         ONLINE       0     0     0
          raidz1    ONLINE       0     0     0
            da0p2   ONLINE       0     0     0
            da1p2   ONLINE       0     0     0
            da2p2   ONLINE       0     0     0
            da3p2   ONLINE       0     0     0

errors: No known data errors


[root@fnas1] /# zpool list
NAME   SIZE   USED  AVAIL    CAP  HEALTH  ALTROOT
zp1   5.44T   153K  5.44T     0%  ONLINE  -
[root@fnas1] /#

so now I created with gui a zpool and added cache with command line

Code:

[root@fnas1] /# zpool add zp1 cache da0

[root@fnas1] /# zpool list
NAME   SIZE   USED  AVAIL    CAP  HEALTH  ALTROOT
zp1   5.44T   153K  5.44T     0%  ONLINE  -
[root@fnas1] /#
[root@fnas1] /# zpool destroy -f zp1
[root@fnas1] /# zpool list
NAME   SIZE   USED  AVAIL    CAP  HEALTH  ALTROOT
zp1    468G   156K   468G     0%  ONLINE  /mnt

[root@fnas1] /# zpool status
  pool: zp1
 state: ONLINE
 scrub: none requested
config:

        NAME        STATE     READ WRITE CKSUM
        zp1         ONLINE       0     0     0
          raidz1    ONLINE       0     0     0
            da1p2   ONLINE       0     0     0
            da2p2   ONLINE       0     0     0
            da3p2   ONLINE       0     0     0
            da4p2   ONLINE       0     0     0
            da5p2   ONLINE       0     0     0
            da6p2   ONLINE       0     0     0

errors: No known data errors
[root@fnas1] /# zpool add zp1 cache da0
[root@fnas1] /# zpool status
  pool: zp1
 state: ONLINE
 scrub: none requested
config:

        NAME        STATE     READ WRITE CKSUM
        zp1         ONLINE       0     0     0
          raidz1    ONLINE       0     0     0
            da1p2   ONLINE       0     0     0
            da2p2   ONLINE       0     0     0
            da3p2   ONLINE       0     0     0
            da4p2   ONLINE       0     0     0
            da5p2   ONLINE       0     0     0
            da6p2   ONLINE       0     0     0
        cache
          da0       ONLINE       0     0     0

errors: No known data errors

that means for me, that the mark "cache" in the gui not work correctly.
adding the cache with command line works well, but the da0 is of course not shown in the disks-area of the gui
but as I think no more tasks with the cache-ssd will be necessary (eg replace or so)

if not yet it should be a ticket to correct the command performed by gui to:

zpool create zp1 raidz da1 da2 ... cache da0

I did not try yet the function of "Spare"

Greetings

uwe

 

[William Grzybowski]

What browser are you using?

I'm pretty sure marking as cache works, at least it did on my test...

Anyway, if you export it from gui and auto-import back it would fix it...

 

[propeller]

I use firefox 8.0
yes, than I don't know what it could be

but before my 'solution' with adding the cache after creating the pool I tested with the LSI-Bios, because I put in 2 SSD disks (no matter, if it make sense) and see only one in the gui. I gues the command line will not show more if I knew the commands ;)
perhaps there is a problem with the LSI-Cards, Samsung SSDs an freenas.
But I didn't take the time to play with other OS like WinServer8 after I failed with VMware and Freenas (Freenas no more booting to end after pass through LSI-PCI-Devices to Freenas-VM),

Thank you

 

[dannyb78]

it's stupid but... freenas is not installed into da0 (missing SSD), right?

 

[propeller]

no, Freenas 8.02 is installed on a 4 GByte Transcend cf-card connected by sata-cf-adapter to one sata-port of the main-board (3 Gbit Sata), because I got USB not to boot (as with an old DELL Power Edge Server from year 2004 which is now gone, but as not with my HP-Micro-Server at home which boots Freenas very well from USB). 8 WD 2 TByte are connected to one LSI-sas/sata-Controller, 4 WD 2 TByte and the 2 SSDs to the other LSI-Controller . I tried several other configurations: connect one SSD to the first SATA-controller, the other to the second, then only one SSD, then changed the SSDs in all these configurations.
May be the BIOS of the mainboard is in conflict with the LSI-BIOS, because when I mark the controllers as "Boot by BIOS" or "Boot by BIOS or OS", then the second SSD is not correct displayed in the mainboard-Bios.
I will leave it for now with one SSD.
Thank you

 

[gcooper]

Thank you for reply.
The version is 8.02
my (zpool) status is healthy ;)

after destroying and create it with gui (marked da0 as "cache") it looks so:

Code:

[root@fnas1] ~# zpool list
NAME   SIZE   USED  AVAIL    CAP  HEALTH  ALTROOT
zp1    476G   138K   476G     0%  ONLINE  -

on command line it looks so:

Code:

[root@fnas1] ~# zpool create zp1 raidz da0 da1 da2 da3
invalid vdev specification
use '-f' to override the following errors:
raidz contains devices of different sizes

That command won't work. This will:

Code:

zpool create zp1 raidz da1 da2 da3 cache da0

Considering that no one has yet reported any issues with this, I'm a bit skeptical of your analysis.

 

[propeller]

correct

yes, that is correct, I didn't delete all my errors from the putty-copy into the posting. sorry
but at the end I wrote the correct command I think.
regards