OpenSolaris derived ZFS NAS/ SAN (OmniOS, OpenIndiana, Solaris and napp-it)

Regarding security, im not sure how to proceed with binding napp-it to an ip.
Server is as follows;

Code:
esxi5 ┬ vSwitch0 (nic1) ┬ port group/192.168.1.254 (esxi management)
      │                 └ port group/192.168.1.201 (OI vm - napp-it management)
      └ vSwitch1 (nic2) ─ port group/192.168.1.200 (OI vm)

Ive disabled nwam and enabled physical. Both adapters in OI have been configured with static ip's (as above)

How do I bind mini-httpd to 192.168.1.201?
I want the napp-it management to only be available from one other pc on the network.
Where do I implement the firewall (on OI, esxi, or the router)? Which ports need to be open - ei just port 81?

i´d like to hear it too. i have the same problem setting up security with IO en esxi
 
Yes, I have a separate job that is set up for my status to run once a week. The alert job shows as this:

email alert send to <MyEmailHere> send by <MySMTPServerHere> every every every every 1317566205 active 23.oct 17:26 - - run now delete

The message I get is as follows:

Subject: *** napp-it ERROR-Alert *** on <MyZFSServerHere>&#8207;
Body: Alert/ Error on <MyZFSServerHere>&#8207; from 23.10.2011 10:57

-disk errors: none

and at the bottom it has the results from a zpool list command.

Any ideas on this one? Is there a config file for the job that I could look at?
 
Regarding security, im not sure how to proceed with binding napp-it to an ip.
Server is as follows;

Code:
esxi5 &#9516; vSwitch0 (nic1) &#9516; port group/192.168.1.254 (esxi management)
      &#9474;                 &#9492; port group/192.168.1.201 (OI vm - napp-it management)
      &#9492; vSwitch1 (nic2) &#9472; port group/192.168.1.200 (OI vm)

Ive disabled nwam and enabled physical. Both adapters in OI have been configured with static ip's (as above)

How do I bind mini-httpd to 192.168.1.201?
I want the napp-it management to only be available from one other pc on the network.
Where do I implement the firewall (on OI, esxi, or the router)? Which ports need to be open - ei just port 81?

why do you have the same subnet of two different vswitches?
 
anyone able to achieve read/write speed constant at around 100MB/s ?

burst at 120+....after a second drops to 70+ then fluctuate ard 70-80+

Any idea on how to improve?
 
How do I bind mini-httpd to 192.168.1.201?
I want the napp-it management to only be available from one other pc on the network.
Where do I implement the firewall (on OI, esxi, or the router)? Which ports need to be open - ei just port 81?

your main option: set a firewall rule that allows port 81 only for grouped appliances
and your management computer. Other ports are only used for replication.

other options:
use zones or another webserver
 
anyone able to achieve read/write speed constant at around 100MB/s ?

burst at 120+....after a second drops to 70+ then fluctuate ard 70-80+

Any idea on how to improve?

its usually due to caches, delayed async writes and copy on write

how to improve:
look for example at napp-it menu system statistic for bottlenecks
otherwise the usual: use more RAM, faster disks, ssd cache and as much vdevs as possible.
 
its usually due to caches, delayed async writes and copy on write

how to improve:
look for example at napp-it menu system statistic for bottlenecks
otherwise the usual: use more RAM, faster disks, ssd cache and as much vdevs as possible.

under benchmark if i can get seq-write 108mb/s and seq-read 176mb/s

i copy files from nas to my windows 7 pc only achieve 80mb/s... isit due to my network switch problem?
 
I ended up with a $50 off to HP SMB and bought a N40L HP Microserver. Since hard drives shot up in price, I ended up sticking four older 1.5tb Seagate 7.2k rpm drives in it until the prices drop. I installed a small 250gb seagate laptop drive right below the 5 1/4 bay for the OS via the onboard ODD sata port. After testing several OS's, I ended up going with OpenIndiana 151 and Napp-it.



Uploaded with ImageShack.us
(4x 1.5tb in Raid Z, sporting 4tb available.)

NexantaStor loaded fine, but I did not like the interface and it had pitifully slow (45mb/s) transfers. OpenSolaris did not want to detect any drives, and running it on top of ESXi was just painfully slow. I never did get around to testing FreeNAS, but so far I am very happy with the OI+ Napp-it combo. This will get left at my folks house an hour away to serve as a offsite backup. Initial tests on the onboard NIC are showing 100mb/s on some 3.5gb iso's which is plenty for a small nas tied to a 30mg comcast line. Hopefully I can run several other services on OI or through Napp-it (apache, ftp, ect.)

I was worried at first after reading several users feedback on slow transfers (40-60mb/s), but this will work fine for my needs.
 
What are the disk space requirements for OpenIndiana + Nappit, can you use a 4-8GB USB drive?
 
Last night i reloaded my OE11 OS after exporting the pool but when i imported it back into the new OS it see's all the shares on there but when i try to access them, it says i don't have the correct premissions? Guest access is enabled.

Any ideas? if you need more info then ill grab it.
 
Last night i reloaded my OE11 OS after exporting the pool but when i imported it back into the new OS it see's all the shares on there but when i try to access them, it says i don't have the correct premissions? Guest access is enabled.

Any ideas? if you need more info then ill grab it.


ontop of this; i can create new files / folders but can't access the old data?
 
If you can create files but not access old data. I would stop all such activities at once. Maybe you are overwriting important parts of the disk? Just focus on getting the data back. google. ask on different forums. etc
 
i would assume its just a permissions thing but as the permissions are set as 755+ for the pool and 777+ for the share - i'm not sure what's gonig on. read only is off
 
heres a pic of the permissions in the zfs folder tab...

zfs_folder.jpg


and here's the error in windows if i try to access a folder - this happens on all of my machines and will allowing me to create new data

error.jpg



obviously this is rather annoying as there is around 4tb's of data on there (ok i dont have a backup) but an export / import of the array is supposed to be easy... right?
 
@ Origin_Unknown: Check the following

- Services > SMB > SMB/CIFS server service (ONLINE)
- Open shell and type "passwd root", put in your root password when prompted and try accessing the shares again.
 
What are the disk space requirements for OpenIndiana + Nappit, can you use a 4-8GB USB drive?

I have read about a few people trying it and getting 18h install times and 45min boot times. I already put mine in to production or I would try it.

I believe a 8gb thumbdrive was used in their testing.
 
heres a pic of the permissions in the zfs folder tab...

zfs_folder.jpg


and here's the error in windows if i try to access a folder - this happens on all of my machines and will allowing me to create new data

error.jpg



obviously this is rather annoying as there is around 4tb's of data on there (ok i dont have a backup) but an export / import of the array is supposed to be easy... right?

how to solve the permission problem
- delete any id-mappings (idmap remove -a)

and/or try
- reset root pw once at console to create a Windows formatted pw
- disable guest access, connect the share as root and reset ACL for ex. to
everyone@=full or modify recursively from Windows or use napp-it ACL extension to do that.

be aware:
Kernel based SMB server is ACL only, just like Windows
 
Last edited:
I have read about a few people trying it and getting 18h install times and 45min boot times. I already put mine in to production or I would try it.

I believe a 8gb thumbdrive was used in their testing.

do not try unless you hhave ssd like stick and usb3,
or it is slow like pain. you shoulöd have about 12 GB+ for a live installe
or 8 GB for a text install (i would prever live with time-slider)

buy a cheap ssd or 2,5" laptop disk instead and connect to Sata (or use a old ide disk)
 
@ Origin_Unknown: Check the following

- Services > SMB > SMB/CIFS server service (ONLINE)
- Open shell and type "passwd root", put in your root password when prompted and try accessing the shares again.

i get a 'Permission denied' from napp-it > console

how to solve the permission problem
- delete any id-mappings (idmap remove -a)

and/or try
- reset root pw once at console to create a Windows formatted pw
- connect the share as root and reset ACL for ex. to
everyone@=full or modify recursively from Windows or use napp-it ACL extension to do that.

be aware:
Kernel based SMB server is ACL only, just like Windows

same as above when changing password
unfortunatly i dont know how to do the other part. - i guess im screwed then?
 
i get a 'Permission denied' from napp-it > console

you need full permissions to change root pw
either login as root at real console or remote via putty (enable with napp-it menu services ssh) or enter at console
sudo passwd root

same as above when changing password
unfortunatly i dont know how to do the other part. - i guess im screwed then?

sudo idmap remove -a

ps
you can enter simple cli commands ike idmap through the napp-it console cmd in the third menu line
! passwd is working only in a real shell because its interactive
(napp-it can handle this for regular user pw, not for root)
 
do not try unless you hhave ssd like stick and usb3,
or it is slow like pain. you shoulöd have about 12 GB+ for a live installe
or 8 GB for a text install (i would prever live with time-slider)

buy a cheap ssd or 2,5" laptop disk instead and connect to Sata (or use a old ide disk)

I only have 4 SATA slots available and want to use those for disks.
 
you need full permissions to change root pw
either login as root at real console or remote via putty (enable with napp-it menu services ssh) or enter at console

sudo passwd root



sudo idmap remove -a

ps
you can enter simple cli commands ike idmap through the napp-it console cmd in the third menu line
! passwd is working only in a real shell because its interactive
(napp-it can handle this for regular user pw, not for root)

cheers for that Gea, i managed to change the password via the real console (on the machine) and then did the 'sudo idmap remove -a' but still can't get into my files.


i don't actually know how to login as the root user via windows to chance the permissions so other than that am i looking at the data being lost?
 
I do believe that i've managed to sort it! *fingers crossed*
i couldn't figure out how to get windows to connect to the share as another user then i remembered the 'connect to server' setting in Ubuntu- after installing that i managed to login as root and it would allow me to see the (still there) files, although it still would not allow me to change any permissions :(
i made a 'test' folder in there Root of the share and copied a few files to it then went into windows and all of a sudden when trying to access the share i got a pop-up asking for a username and password - when i entered root & the password it has allowed me to change the file permissions - which windows is happily changing at the moment (its going to take a while, one directory has over 100k files / folders in)

hopefully that means i can access the files properly again! and guest access i assume will be restored?
 
cheers for that Gea, i managed to change the password via the real console (on the machine) and then did the 'sudo idmap remove -a' but still can't get into my files.


i don't actually know how to login as the root user via windows to chance the permissions so other than that am i looking at the data being lost?

you can access your data:
from console with midnight commander: sudo mc
from nautilus file browser if you have installed live version: sudo nautilus
from windows: unshare the folder, reshare it without guest access,
on next access, you need to enter username (root) and pw and set acl for everyone@, then unshare and reshare with guest access
 
I only have 4 SATA slots available and want to use those for disks.

use 3 bigger data disks
use extra controller
use another mainboard with 6 sata
use another os (free-bsd with os on datapool),
- but its also not recomended there

i would prefer option 2
 
I have napp-it 0.500s on top of Solaris 11 Express. I had previously tested the email alerts and they worked exactly as I expected. However since then my SMTP server changed. At first I didn't see where I could change it, so I tried deleting the job and recreating it. Now instead of only alerting me when there is an error, I get an email every day saying there are no errors. Did I do something wrong and how can I get it back to just alerting if there is an error?

which napp-it version?
if older than 0.5s, please try an update and recreate alert job

It was on 0.5s when I deleted and recreated it. ;)
 
I have done some more testing regarding AES-NI acceleration and I can certainly rule out ESXi as the problem now. I did a new install of Solaris 11 Express directly on the hardware and found that performance was still poor (although better than with ESXi, mainly due to more cores).

This made me doubt that ZFS encryption supports AES-NI at all, so I decided to turn off AES-NI in the BIOS and see if that made any difference. It turns out that it does not make much difference at all.

Using bonnie++ for the benchmarks now, first on an unencrypted folder:
Code:
Version 1.03c       ------Sequential Output------ --Sequential Input- --Random-
                    -Per Chr- --Block-- -Rewrite- -Per Chr- --Block-- --Seeks--
Machine        Size K/sec %CP K/sec %CP K/sec %CP K/sec %CP K/sec %CP  /sec %CP
solaris         16G 118768  97 189487  15 89507  11 118534  89 197236   9 563.7   1
                    ------Sequential Create------ --------Random Create--------
                    -Create-- --Read--- -Delete-- -Create-- --Read--- -Delete--
              files  /sec %CP  /sec %CP  /sec %CP  /sec %CP  /sec %CP  /sec %CP
                 16 +++++ +++ +++++ +++ +++++ +++ +++++ +++ +++++ +++ 30524 100

Then on an encrypted folder with AES-NI on:
Code:
Version 1.03c       ------Sequential Output------ --Sequential Input- --Random-
                    -Per Chr- --Block-- -Rewrite- -Per Chr- --Block-- --Seeks--
Machine        Size K/sec %CP K/sec %CP K/sec %CP K/sec %CP K/sec %CP  /sec %CP
solaris         16G 95504  81 76644   8 53215   8 90612  97 197748   9 477.3   0
                    ------Sequential Create------ --------Random Create--------
                    -Create-- --Read--- -Delete-- -Create-- --Read--- -Delete--
              files  /sec %CP  /sec %CP  /sec %CP  /sec %CP  /sec %CP  /sec %CP
                 16 31769  97 +++++ +++ +++++ +++ 18941  98 +++++ +++ +++++ +++

And finally the encrypted folder with AES-NI off:
Code:
Version 1.03c       ------Sequential Output------ --Sequential Input- --Random-
                    -Per Chr- --Block-- -Rewrite- -Per Chr- --Block-- --Seeks--
Machine        Size K/sec %CP K/sec %CP K/sec %CP K/sec %CP K/sec %CP  /sec %CP
solaris         16G 79832  72 78021   9 46583   7 104403  95 170061   8 554.3   1
                    ------Sequential Create------ --------Random Create--------
                    -Create-- --Read--- -Delete-- -Create-- --Read--- -Delete--
              files  /sec %CP  /sec %CP  /sec %CP  /sec %CP  /sec %CP  /sec %CP
                 16 +++++ +++ +++++ +++ +++++ +++ +++++ +++ +++++ +++ +++++ +++

It is hard to draw a conclusion by looking at those numbers, because the CPU is just barely the bottleneck here. However, it should be noted that CPU usage was very high when performing the benchmark on the encrypted filesystems regardless of whether AES-NI was turned on or off. I would not expect the CPU load to be so high if AES-NI was indeed being utilized.

Then I figured it would be easier to see if AES-NI is working by going back to ESXi and running Solaris with only 2 vCPUs. That would definitely make the CPU the bottleneck. Here I got some very interesting results.

First with AES-NI on:
Code:
Version 1.03c       ------Sequential Output------ --Sequential Input- --Random-
                    -Per Chr- --Block-- -Rewrite- -Per Chr- --Block-- --Seeks--
Machine        Size K/sec %CP K/sec %CP K/sec %CP K/sec %CP K/sec %CP  /sec %CP
solaris1        10G 41791  30 39346   3 21610   2 60120  43 79505   3  1365   2
                    ------Sequential Create------ --------Random Create--------
                    -Create-- --Read--- -Delete-- -Create-- --Read--- -Delete--
              files  /sec %CP  /sec %CP  /sec %CP  /sec %CP  /sec %CP  /sec %CP
                 16 +++++ +++ +++++ +++ +++++ +++ 32295  99 +++++ +++ +++++ +++

AES-NI off:
Code:
Version 1.03c       ------Sequential Output------ --Sequential Input- --Random-
                    -Per Chr- --Block-- -Rewrite- -Per Chr- --Block-- --Seeks--
Machine        Size K/sec %CP K/sec %CP K/sec %CP K/sec %CP K/sec %CP  /sec %CP
solaris1        10G 49511  35 53982   4 23716   2 80445  56 100416   3 470.4   0
                    ------Sequential Create------ --------Random Create--------
                    -Create-- --Read--- -Delete-- -Create-- --Read--- -Delete--
              files  /sec %CP  /sec %CP  /sec %CP  /sec %CP  /sec %CP  /sec %CP
                 16 27401  72 +++++ +++ +++++ +++ +++++ +++ +++++ +++ +++++ +++

There appears to be a difference and it's probably not what you expected. Performance is worse with AES-NI enabled! Although the results are somewhat inconsistent, I think it's fairly safe to say that AES-NI is not working with ZFS encryption in Solaris.

Now, I'm curious to hear if anyone actually is getting AES-NI acceleration with ZFS using different hardware. Is this just an issue with Xeon E3?
 
If you compile and run this it will tell you if you are getting AES-NI support properly down the entire stack (though not neccecarily if zfs encryption is working fine)

Code:
#include <stdio.h>
#include <sys/auxv.h>

void main(void) 
{
     uint_t ui = 0; (void) getisax(&ui, 1);
     printf("AES-NI instructions are %spresent.\\n", (ui & AV_386_AES) ? "" : "not ");
}
 
If you compile and run this it will tell you if you are getting AES-NI support properly down the entire stack (though not neccecarily if zfs encryption is working fine)

Code:
#include <stdio.h>
#include <sys/auxv.h>

void main(void) 
{
     uint_t ui = 0; (void) getisax(&ui, 1);
     printf("AES-NI instructions are %spresent.\\n", (ui & AV_386_AES) ? "" : "not ");
}
Yes, AES-NI is seen by Solaris but it is not used by ZFS. I also found an easier way to get this information; just type isainfo -v and look for 'aes'.

Have you tried AES 128 vs 256?

Yes, it's pretty much the same thing. My initial tests were with AES 256, but I also tested 128 on my Crucial M4 (basic vdev) with AES-NI enabled:

Unencrypted:
Code:
Version 1.03c       ------Sequential Output------ --Sequential Input- --Random-
                    -Per Chr- --Block-- -Rewrite- -Per Chr- --Block-- --Seeks--
Machine        Size K/sec %CP K/sec %CP K/sec %CP K/sec %CP K/sec %CP  /sec %CP
solaris1        10G 98010  71 78492   6 61398   7 134431  97 506835  20  8448  14
                    ------Sequential Create------ --------Random Create--------
                    -Create-- --Read--- -Delete-- -Create-- --Read--- -Delete--
              files  /sec %CP  /sec %CP  /sec %CP  /sec %CP  /sec %CP  /sec %CP
                 16 +++++ +++ +++++ +++ +++++ +++ +++++ +++ +++++ +++ +++++ +++

Encrypted (aes-128-ccm):
Code:
Version 1.03c       ------Sequential Output------ --Sequential Input- --Random-
                    -Per Chr- --Block-- -Rewrite- -Per Chr- --Block-- --Seeks--
Machine        Size K/sec %CP K/sec %CP K/sec %CP K/sec %CP K/sec %CP  /sec %CP
solaris1        10G 47964  35 58297   4 32467   3 72478  52 89766   2  1197   1
                    ------Sequential Create------ --------Random Create--------
                    -Create-- --Read--- -Delete-- -Create-- --Read--- -Delete--
              files  /sec %CP  /sec %CP  /sec %CP  /sec %CP  /sec %CP  /sec %CP
                 16 +++++ +++ +++++ +++ +++++ +++ 27107  78 +++++ +++ +++++ +++
 
I'm assuming there is a really simple explanation for the problem I'm having, but I've been banging my head against it for too long, so I thought I'd check here.

I have a great all-in-one napp-it machine up and running. Windows file sharing is great. I recently tried to get a machine running fedora to mount via nfs a napp-it shared ZFS folder.

using the command 'mount 192.168.0.90:/music /tmp/music' yields:
'access denied by server while mounting 192.168.0.90:/music'

I am logged into the fedora machine with a unix account that has administrative access to the ZFS solairis machine. the NFS share is enabled in the napp-it GUI.

What am I doing wrong? Do I need to edit /etc/exports, hosts.allow, hosts.deny? I assumed the napp-it gui would do that for me... Your help is greatly appreciated. I have to walk away from this one for a few minutes. It seems like a trivial problem, but I cannot make it work.
 
Yo,

Am new here so bear with me....
I come from a retail NAS environement and I am looking to build my first server.
Problem with NAS they are expensive and once all trays are full then you need to buy a new one so I ended up with 4 machines.
Like I said I'm a newbie when it comes to building my own server and the huge amount of choice doesn't make it any better. I'm currently testing Unraid but the write speeds are a dissapointment to me especially when it come to copying 12 to 20tb (33mb/s with parity calculation)
So I have been reading about ZFS but there are so many different choices! Basically I wanna have redundancy (minimal raid-5 or Raid-z) and maximum space available! I also read about Freenas which also has ZFS so it's hard for me to know all the differences pros/cons.
Basically I want to store my movie collection which consistes of BD-iso and MKV, also music, family pictures and videos. The server doesn't need to be on 24/7, only when I require it. It has to be able to stream to my media players using NFS or SMB and stream to my ipad.
So here is the hardware I allready purchased:

Norco RPC-4224 v3
Supermicro X9SCM-F server board with IPMI 2.0
Intel Xeon E-1230
2x4 GB Kingston ECC unbuffered 1333mhz
Seasonic X-760 80+ Gold PSU
Supermicro AOC SAS2LP MV8 sata-600 8-port HBA (don't think it works with ZFS)
4x Hitachi 5K3000 2TB sata-600 5400rpm
4x WD EARX 2TB sata-600 5400rpm
3 OCZ Vertex 3 sata-600 SSD 120GB (use them as write/read cache?)
I have 1 Hitachi 7K3000 3TB 7200rpm Sata-600 and 1 WD30EZRX 5400rpm/3tB/sata-600 I can use


I'm planning to buy 2 or 3
IBM ServeRAID M1015 SAS RAID FRU 46M0861 LSI SAS9220-8i and flash them into IT mode as they should work in ZFS. If I have 3 I can use all 24 cages for storage (3x8) and I ordered a Scythe Kama Stay to mount the 3 SSD's inside the Norco and cool the HBA cards at the same time!
http://www.scythe-usa.com/product/acc/071/sckst1000_detail.html
Is it possible to run both let say Opensolaris and Napp-it from 1 USB stick?
How do I attach the drives I will use in a pool, one pool/one HBA or different HBA's?
How do I use the cache/SSD's?
What speeds will I be expecting? I run Cat-6 in my house.
Does ZFS tell me when a drive fails? Can I spin down the drives/pools which are not in use as I am planning to build 3 pools?

I hope someone can answer all this as it's a jungle out there and so many options...

gr33tz
 
I'm assuming there is a really simple explanation for the problem I'm having, but I've been banging my head against it for too long, so I thought I'd check here.

I have a great all-in-one napp-it machine up and running. Windows file sharing is great. I recently tried to get a machine running fedora to mount via nfs a napp-it shared ZFS folder.

using the command 'mount 192.168.0.90:/music /tmp/music' yields:
'access denied by server while mounting 192.168.0.90:/music'

I am logged into the fedora machine with a unix account that has administrative access to the ZFS solairis machine. the NFS share is enabled in the napp-it GUI.

What am I doing wrong? Do I need to edit /etc/exports, hosts.allow, hosts.deny? I assumed the napp-it gui would do that for me... Your help is greatly appreciated. I have to walk away from this one for a few minutes. It seems like a trivial problem, but I cannot make it work.

i suppose, its just a permission problem.
allow everyone access to shared files

ex from windows: smb-login as root and set everyone = modify recursively

ps
napp-it just enables nfs-sharing, you must set permissions accordingly to you needs
either as file/folder permissions or as parameter in napp-it's nfs sharing menu.
 
Yo,

Am new here so bear with me....
I come from a retail NAS environement and I am looking to build my first server.
Problem with NAS they are expensive and once all trays are full then you need to buy a new one so I ended up with 4 machines.
Like I said I'm a newbie when it comes to building my own server and the huge amount of choice doesn't make it any better. I'm currently testing Unraid but the write speeds are a dissapointment to me especially when it come to copying 12 to 20tb (33mb/s with parity calculation)
So I have been reading about ZFS but there are so many different choices! Basically I wanna have redundancy (minimal raid-5 or Raid-z) and maximum space available! I also read about Freenas which also has ZFS so it's hard for me to know all the differences pros/cons.
Basically I want to store my movie collection which consistes of BD-iso and MKV, also music, family pictures and videos. The server doesn't need to be on 24/7, only when I require it. It has to be able to stream to my media players using NFS or SMB and stream to my ipad.
So here is the hardware I allready purchased:

Norco RPC-4224 v3
Supermicro X9SCM-F server board with IPMI 2.0
Intel Xeon E-1230
2x4 GB Kingston ECC unbuffered 1333mhz
Seasonic X-760 80+ Gold PSU
Supermicro AOC SAS2LP MV8 sata-600 8-port HBA (don't think it works with ZFS)
4x Hitachi 5K3000 2TB sata-600 5400rpm
4x WD EARX 2TB sata-600 5400rpm
3 OCZ Vertex 3 sata-600 SSD 120GB (use them as write/read cache?)
I have 1 Hitachi 7K3000 3TB 7200rpm Sata-600 and 1 WD30EZRX 5400rpm/3tB/sata-600 I can use


I'm planning to buy 2 or 3
IBM ServeRAID M1015 SAS RAID FRU 46M0861 LSI SAS9220-8i and flash them into IT mode as they should work in ZFS. If I have 3 I can use all 24 cages for storage (3x8) and I ordered a Scythe Kama Stay to mount the 3 SSD's inside the Norco and cool the HBA cards at the same time!
http://www.scythe-usa.com/product/acc/071/sckst1000_detail.html
Is it possible to run both let say Opensolaris and Napp-it from 1 USB stick?
How do I attach the drives I will use in a pool, one pool/one HBA or different HBA's?
How do I use the cache/SSD's?
What speeds will I be expecting? I run Cat-6 in my house.
Does ZFS tell me when a drive fails? Can I spin down the drives/pools which are not in use as I am planning to build 3 pools?

I hope someone can answer all this as it's a jungle out there and so many options...

gr33tz

- my best to use SAS controller is a LSI 9211
best cheap nearly identical is a SuperMicro AOC-USAS2-L8e (wrong side mounted)

If you can get it cheap, the IBM is an option after reflashing
new LSI SAS2 based expanders like RES2SV240 are also quite well on a 2008 SAS controller

-OI on a usb stick is a nogo
use a 2,5" laptop drive or a cheap small sata ssd

- it does not matter where your disk is attached.
you can build pools from any disks on any controller

-hybrid storage:
just add a ssd disk as read or write cache to a pool

speed:
depends on your raid-level and I/O needs
mirror=fast
raid-z=capacity
with media files on Gbe: enough

spin down:
you can spindown disks (poolwise)
or shutdown via timer, power on via ipmi

notify:
yes, example via napp-it - send email alerts
 
Last edited:
Yo,

Thanks for the info....so what version of ZFS should I install cuz there are so many versions! I was planning to go for Oracle Solaris Express 11 Live edition.... I also see that Express 1 and Open Indiana have USB editions so it should be possible to boot from a USB disk ,right?

thanks
 
Last edited:
I believe those USB editions are actually Live versions meant for maintenance and installing. I don't think you can use them on a real system, AFAIK.
 
I would go with OI for the ability to get patches/bugfixes (unless you want to pay for an oracle support contract).

I tried SE11 & OI on a USB drive - it works, but performance is terrible (8hour installs, 1-2 hour boots). Just pick up any random 2.5" laptop hard drive (or a pair and mirror them) - doesn't have to be fancy at all. Should be < 50 each (well, don't know with the current HD price hikes)
 
Back
Top