[OpenIndiana-discuss] Low low end server

Alex Smith (K4RNT) shadowhunter at gmail.com
Sun Feb 9 21:35:28 UTC 2014


I don't know if it affects your setup, but I thought the 9211 was 6Gbps SAS
only, and 3Gbps SATA. Are your disks SAS? I'm not familiar with any drives
that WD makes being SAS, but maybe it was just the environment I was in
didn't use them.

" ' With the first link, the chain is forged. The first speech censured,
the first thought forbidden, the first freedom denied, chains us all
irrevocably.' Those words were uttered by Judge Aaron Satie as wisdom and
warning... The first time any man's freedom is trodden on we’re all
damaged." - Jean-Luc Picard, quoting Judge Aaron Satie, Star Trek: TNG
episode "The Drumhead"
- Alex Smith
- Dulles Technology Corridor (Chantilly/Ashburn/Dulles), Virginia USA


On Sun, Feb 9, 2014 at 4:19 PM, Jim Klimov <jimklimov at cos.ru> wrote:

> Sorry about the unclarities, I was away from both access to that machine
> and our shopping spec :)
>
> Ultimately, what my brother shopped for was:
> * Chassis - minimal N54L with 2Gb RAM and 250Gb HDD which went elsewhere
> * RAM - 2 * Kingston / KVR1333D3E9S/8G
>   2*8Gb ECC (9*1024Mbyte = 72*1Gbit) DDR3 PC3-10600E-9 NOT_REGISTERED
>   (UNBUFFERED)
> * data pool HDDs - 4*4Tb
>   initially a set of Hitachi Deskstar 7K4000: H3IK40003272SE but they
>   happened to be a bad batch (2 did not even start, one was very slow,
>   just one seemed okay); replaced with
>   a set of WD Red 4Tb ("<ATA-WDC WD40EFRX-68W-0A80-3.64TB>" in format)
> * OS/cache SSDs - 2*Seagate Enterprise SSD 120Gb ST120FN0021
> * a 4*2.5" disk rack to fit into the 5.25" ODD slot
>   (Thermaltake ST0046Z Max5 Quad)
> * the controller is apparently an original LSI 9211-8i
>   when picking your parts, make sure that the card is small
>   (half-length/half-height), we first tried a cheaper equivalent
>   which did not fit into the cramped little box
> * an additional 1*SFF8087-4*SATA cable (overall two cables, one provided
>   with the server, none provided with this controller purchase (not a
>   "kit" bundle)); make sure cables are not too long nor too short
> * an USB flash drive for the OS image for installation :)
>
> apparently, that was it
>
>
> On 2014-02-09 11:48, Hans J. Albertsson wrote:
>
>> 2. does not use stock SATA ports at all,
>>>
>> Why? Are these SATA ports bad or substandard or what? Are you not even
>> using them for booting?
>>
>
> On one hand, the Mobo SATA controller is 3Gbps and the LSI card has
> 8*6Gbps ports; on another - this was the configuration that my brother
> tried first, it worked, and we avoided changes "from good to best".
> Ain't broke - don't fix, and all that :)
>
> One of the SFF8087 connectors goes to the main 4-disk bay, another
> goes to the add-on 4*2.5" bay in the ODD slot. There is not really
> much space to use more cabling than necessary.
>
> Besides, the controller is deemed to be a more "intellectual" piece
> of hardware and I can poke it with lsiutil, etc., unlike the Mobo ports.
>
>
>  4. 4*4TB disks for the data pool (raidz1, though I'd rather have more
>>> disks and a
>>> higher redundancy) and
>>>
>> What disk type?
>> With 8 ports on the LSI controller, I suppose you could add an external
>> ICY Box for 4 more disks? If you boot from the on-bord SATA ports.
>>
>
> This particular controller is 8i - 8 internal ports (over two 4-port
> connectors, not 8 SATA/SAS connectors). I guess it would make more
> sense for redundancy to use the Mobo eSATA port for external disks
> (spread the load over two controllers, especially if using mirroring).
>
>
>
>>
>>  5. 2*120Gb Samsung Pro SSDs (with powerloss
>>> protection, formatted to use 100Gb for mirrored rpool/mirrored
>>> zil/striped l2arc since their 100Gb sub-model has much higher
>>> reliability and speed - but is not on sale here). I'd say that the
>>> L2ARC on this box could use more volume, it (~2*60Gb) fills up
>>> pretty quickly and does not place a big toll on RAM yet. Disk IO
>>> performance is pretty decent for a home gigabit LAN, processing
>>> not so much (i.e. for compilation), although still quite usable :)
>>>
>>>  The above is a wee bit unclear! Please clarify.
>>
>
>
> The disks are partitioned as follows:
>
> root at n54l:~# parted /dev/dsk/c3t5000C5002FF029D5d0p0 pri
> Model: Generic Ide (ide)
> Disk /dev/dsk/c3t5000C5002FF029D5d0p0: 120GB
> Sector size (logical/physical): 512B/512B
> Partition Table: msdos
>
> Number  Start   End     Size    Type     File system  Flags
>  1      8354kB  33.7GB  33.7GB  primary  solaris      boot
>  2      33.7GB  39.2GB  5417MB  primary
>  3      39.2GB  100GB   61.1GB  primary
>
>
> This uses up 100Gb of the 120Gb available, in the hopes that the
> drive would perform similar to the stock 100Gb model with much
> better characteristics.
>
> These partitions are used for rpool, and for data pool's zil and
> l2arc, respectively:
>
> root at n54l:~# zpool status -v
>   pool: pool
>  state: ONLINE
>   scan: scrub repaired 0 in 13h11m with 0 errors on Mon Jan 27 13:16:43
> 2014
> config:
>
>         NAME                         STATE     READ WRITE CKSUM
>         pool                         ONLINE       0     0     0
>           raidz1-0                   ONLINE       0     0     0
>             c3t50014EE2B3C27E80d0    ONLINE       0     0     0
>             c3t50014EE2B3C2749Bd0    ONLINE       0     0     0
>             c3t50014EE25E6CE061d0    ONLINE       0     0     0
>             c3t50014EE20917A8B4d0    ONLINE       0     0     0
>         logs
>           mirror-1                   ONLINE       0     0     0
>             c3t5000C5002FF029D5d0p2  ONLINE       0     0     0
>             c3t5000C5002FF02451d0p2  ONLINE       0     0     0
>         cache
>           c3t5000C5002FF029D5d0p3    ONLINE       0     0     0
>           c3t5000C5002FF02451d0p3    ONLINE       0     0     0
>
> errors: No known data errors
>
>   pool: rpool
>  state: ONLINE
>   scan: scrub repaired 0 in 0h1m with 0 errors on Mon Jan 27 13:19:07 2014
> config:
>
>         NAME                         STATE     READ WRITE CKSUM
>         rpool                        ONLINE       0     0     0
>           mirror-0                   ONLINE       0     0     0
>             c3t5000C5002FF029D5d0s0  ONLINE       0     0     0
>             c3t5000C5002FF02451d0s0  ONLINE       0     0     0
>
> errors: No known data errors
>
> HTH,
> //Jim Klimov
>
>
>
> _______________________________________________
> OpenIndiana-discuss mailing list
> OpenIndiana-discuss at openindiana.org
> http://openindiana.org/mailman/listinfo/openindiana-discuss
>


More information about the OpenIndiana-discuss mailing list