ZFS and SAS expanders with SATA drives a toxic combo?

madrebel said:
no license required.
Click to expand...

If you want the latest you do - otherwise it's OpenSolaris b134 (and it's open source derivatives such as OpenIndiana, Nexenta etc) or the FreeBSD or Linux ports.

At some point, there will be a divergence from Oracle's ZFS and the open ZFS forks - either that or the open ZFS fork will stay still!


madrebel said:
however you can't just turn the ship around. emc, netapp, etc have billions invested in their platforms. further it would be impossible for them to justify the insane premium using an open source base.
Click to expand...

Perhaps....but nobody is actually forced to buy high end storage platforms etc etc.



madrebel said:
this is the same thing emc/netapp do only they charge an insane premium.
Click to expand...

I see....so if I buy a few enterprise drives and shove them in a PC, does that make it a high end storage platform? :D

Seriously, I think you know that there is a lot more to it than just that.


madrebel said:
so performance and availability do not qualify as a high end solution .... unless it says emc or netapp and you spend over a million dollars for it?
Click to expand...

They are certainly part of it, but you'd have to elaborate on what you mean by availability - the targets for high end storage platforms are pretty high!
 
taroumaru said:
Heck even packetboy's post [#99] in this thread, he claims to have seen no data loss from using SATA drives with various SAS expanders & SAS controllers; this too over 1.5 years using 3 different systems, for a combined 4.5 years of usage! I even have the exact same SM chassis as him [well he suggested it when I asked for opinions on SM SC846 almost a year ago; thanks packetboy], although I've only been running for a few months using Solaris 11.11 & not as long as him, no problems so far. I also have this gut feeling, that packetboy is using an OS other than Nexenta! I have already sent him a PM asking him this very same question.
/showpost.php?p=1038560587&postcount=825[/url]
Click to expand...


Code: 
System 1:

# uname -a
SunOS zulu01 5.11 oi_151a4 i86pc i386 i86pc Solaris

Main menu, select an option:  [1-99 or e/p/w or 0 to quit] 1

Current active firmware version is 0d000100 (13.00.01)
Firmware image's version is MPTFW-13.00.01.00-IT
  LSI Logic
  Not Packaged Yet
x86 BIOS image's version is MPT2BIOS-7.25.00.00 (2012.02.17)


Main menu, select an option:  [1-99 or e/p/w or 0 to quit] 16

SAS2008's links are down, down, down, 6.0 G, 6.0 G, 6.0 G, 6.0 G, 6.0 G

 B___T     SASAddress     PhyNum  Handle  Parent  Type
        500605b0021a1b20           0001           SAS Initiator
        500605b0021a1b20           0002           SAS Initiator
        500605b0021a1b20           0003           SAS Initiator
        500605b0021a1b20           0004           SAS Initiator
        500605b0021a1b20           0005           SAS Initiator
        500605b0021a1b20           0006           SAS Initiator
        500605b0021a1b20           0007           SAS Initiator
        500605b0021a1b20           0008           SAS Initiator
 0   9  500110a0014c368c     3     0009    0001   SAS Target
 0  10  500143800547f926     4     000a    0002   Edge Expander
 0  11  500143800547f900     0     000b    000a   SATA Target
 0  12  500143800547f901     1     000c    000a   SATA Target
 0  13  500143800547f902     2     000d    000a   SATA Target
 0  14  500143800547f903     3     000e    000a   SATA Target
 0  15  500143800547f905     5     000f    000a   SATA Target
 0  16  500143800547f906     6     0010    000a   SATA Target
 0  17  500143800547f907     7     0011    000a   SATA Target
 0  18  500143800547f909     9     0012    000a   SATA Target
 0  19  500143800547f90a    10     0013    000a   SATA Target
 0  20  500143800547f90b    11     0014    000a   SATA Target
 0  21  500143800547f910    16     0015    000a   SATA Target
 0  22  500143800547f911    17     0016    000a   SATA Target
 0  23  500143800547f912    18     0017    000a   SATA Target
 0  24  500143800547f913    19     0018    000a   SATA Target
 0  25  500143800547f925    36     0019    000a   SAS Initiator and Target

Type      NumPhys    PhyNum  Handle     PhyNum  Handle  Port  Speed
Adapter      8          3     0001  -->    0     0009     0    6.0
                        4     0002  -->   27     000a     1    6.0
                        5     0002  -->   26     000a     1    6.0
                        6     0002  -->   25     000a     1    6.0
                        7     0002  -->   24     000a     1    6.0

Expander    37          0     000a  -->    0     000b     1    3.0
                        1     000a  -->    0     000c     1    3.0
                        2     000a  -->    0     000d     1    3.0
                        3     000a  -->    0     000e     1    3.0
                        5     000a  -->    0     000f     1    3.0
                        6     000a  -->    0     0010     1    3.0
                        7     000a  -->    0     0011     1    3.0
                        9     000a  -->    0     0012     1    3.0
                       10     000a  -->    0     0013     1    3.0
                       11     000a  -->    0     0014     1    3.0
                       16     000a  -->    0     0015     1    3.0
                       17     000a  -->    0     0016     1    3.0
                       18     000a  -->    0     0017     1    3.0
                       19     000a  -->    0     0018     1    3.0
                       24     000a  -->    7     0002     1    6.0
                       25     000a  -->    6     0002     1    6.0
                       26     000a  -->    5     0002     1    6.0
                       27     000a  -->    4     0002     1    6.0
                       36     000a  -->   36     0019     1    6.0

Enclosure Handle   Slots       SASAddress       B___T (SEP)
           0001      8      500605b0021a1b20
           0002      0      500143800547f925    0  25
 
I bought a 70 drive HP MDS600 a while ago and have been keeping tabs on the whole SATA over SAS stuff, just because I could get so much more capacity for the $$ using SATA drives in this thing.

A couple of weeks ago I borrowed ~40 drives from a friend's previous storage upgrades. This stuff is a mish mash of 2-4 year old hardware that I thought would make a good 'worst case' torture test for ZFS on this box. It's a mixture of ~14 400GB SATA, ~12 750 GB SATA, and ~20 1TB SAS. I threw them all together into the same pool as a couple of raidz2s and wrote a python script to torture test the drives. It essentially run 8 independent threads that randomly write or verify random blocks of data of 1MB-15GB of size. After the drive reaches 90% capacity it also starts randomly deleting files. I've let it run mostly non-stop for a week or so on this thing while occasionally randomly pulling drives or running scrubs. I've even tested pulling one of the SAS cables while it was going.

The only odd problem I ran into was the pool stopped responding after getting a lot of scsi errors. It turns out it they were coming from one of the SAS drives being in a specific slot in the box. I actually swapped that SAS drive with a SATA drive in a different slot and everything was fine. I figure maybe there was just something weird about the backplane in that particular slot.

During that time I've written and read 100TB+ of data on the pool, and had read and write speeds averaging ~200MB/sec up to 400MB/sec.

At what point can I call BS on the whole toxic SATA over SAS thing?
 
You should know that abscence of evidence, is not evidence of abscence. This proves nothing.

But anyway, some say the problem is fixed and maybe it is. I dont know. There are no links confirming that the problem is fixed. We have asked numerous times after more information but noone has succeeded in finding such links.
 
You are right, as a scientist I am familiar with the distinction. Unfortunately I also know that waiting for something to be 'proved' is often anathema to making decisions in the near term. I didn't see much point in foregoing big cost savings just to accept possibly outdated 'conventional wisdom'.
 
You must log in or register to reply here.
Back
Top