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

Freak1

Limp Gawd
Joined
Sep 9, 2009
Messages
191
AD server must also be the dns-server who translates names to ip.
Therefor you must use the IP of your AD server when you join with napp-it

(or if you do it manually, use ip at dns entry)

check also if you ad server has already a host with your solaris hostname,
if so delete it andtry rejoin (try at least twice to join).

Gea

Ok i tried this:
name of domain: freak and freak.local
ad-domain-server: 192.168.1.19

Then the dns looks like this:

search freak.local
domain freak.local
nameserver 192.168.1.19

[libdefaults]
default_realm = FREAK.LOCAL

[realms]
FREAK.LOCAL = {
kdc = 192.168.1.19
admin_server = 192.168.1.19
kpasswd_server = 192.168.1.19
kpasswd_protocol = SET_CHANGE
}

[domain_realm]
.freak.local = FREAK.LOCAL

[logging]
default = FILE:/var/krb5/kdc.log
kdc = FILE:/var/krb5/kdc.log
→ kdc_rotate = {

I tried 2 times with both freak and freak.local i also checked the domain server the openindiana server is not under computers. i still fails... :-(
 

_Gea

2[H]4U
Joined
Dec 5, 2010
Messages
4,065
Ok i tried this:
name of domain: freak and freak.local
ad-domain-server: 192.168.1.19

Then the dns looks like this:

search freak.local
domain freak.local
nameserver 192.168.1.19

[libdefaults]
default_realm = FREAK.LOCAL

[realms]
FREAK.LOCAL = {
kdc = 192.168.1.19
admin_server = 192.168.1.19
kpasswd_server = 192.168.1.19
kpasswd_protocol = SET_CHANGE
}

[domain_realm]
.freak.local = FREAK.LOCAL

[logging]
default = FILE:/var/krb5/kdc.log
kdc = FILE:/var/krb5/kdc.log
→ kdc_rotate = {

I tried 2 times with both freak and freak.local i also checked the domain server the openindiana server is not under computers. i still fails... :-(


a domain is usually organisation.TLD ex myuniversity.org
but it may work with freak.local

beside that i can't see a problem from your entries.

ps:
we reached thread 1000


Gea
 

Freak1

Limp Gawd
Joined
Sep 9, 2009
Messages
191
a domain is usually organisation.TLD ex myuniversity.org
but it may work with freak.local

beside that i can't see a problem from your entries.

ps:
we reached thread 1000


Gea

It should work. I have 3 PCs connected to the domain controller XP, Win7 and 2008r2, all works fine. In windows i type in freak here i tries both freak and freak.local but all fails...

Is there really nothing else i can try?
 

_Gea

2[H]4U
Joined
Dec 5, 2010
Messages
4,065
This is a really newbe question but how do I install the zpool patch that appears here http://wiki.openindiana.org/pages/viewpage.action?pageId=4883847
So as I can format my 4k drives correctly in Napp-it

And another question how do you scrub an existing Raidz pool.....

OpenIndiana is beta, the modified zpool executable is experimental.
Do not use with critical data.

I personally would not do anything and accept a slightly reduced performance.
if you have the choice, just avoid 4k disks.


about scrubbing
goto menu jobs-autoscrub and define a weekly scrubbing
(enable auto 15min)

Gea
 
Last edited:

forumator

Weaksauce
Joined
Nov 6, 2009
Messages
93
This is a really newbe question but how do I install the zpool patch that appears here http://wiki.openindiana.org/pages/viewpage.action?pageId=4883847
So as I can format my 4k drives correctly in Napp-it

I'm curious about this as well, I've got 4K drives and want to make sure I create a proper RAIDZ...or is perhaps the ability to create a patched zpool going to be implemented in napp-it?

Also wondering how important it is to have the "magic" number of disks, ex. 6 disk RAID-Z2, and what the consequences would be to have the wrong number of disks?
 

_Gea

2[H]4U
Joined
Dec 5, 2010
Messages
4,065
There is no proper raid-z layout or magic number of disks with ZFS.
ZFS is designed to do its best with any number of disks in a pool or vdev,
independant of the number of disks or if they seems to be 512bit drives instead of
a real 4k value.

The problem is not with ZFS, the problem is YOU, if you want to get the very best
speed out of a pool or vdev although you bought the "wrong" disks or use the "wrong" number.

From what i have seen, you may have a up to 10-20% performance impact with 4k disks in sequential
performance . I do not care because 4k disks are always the low power and slow ones and i bought them
because they are cheap not fast. With a few disks you are above the speed of Gbit Ethernet, why do you care?

In terms of IO/s cheap disks they are not very good at all and i do not expect a serious difference to non 4k low power and cheap disks.
So who cares?.Use a SSD read cache if you have problems, use a SSD write cache mirror only if really needed.
Buy fast disks or use SSD only pools if you need real power. Do not expect a Porsche from no money.

The same with the number of disks.
The best is a division of 128 for data disks only without parity to have a fully balanced vdev (all data blocks are written or read always from
all disks) but thats not real live. You start with a pool with one or two vdevs and add more vdevs later. ZFS does not
rebalance a pool so only new data is using all disks with full striped performance. Thats the way it works.

My opinion;:
do not care about at all, ZFS will do its best automatically in terms of performance and data security.
Care about only if you want to design a real high speed storage from begin.
In other cases, do not think about the difference, mostly go may get not the very best but always a good enough.

Gea
 
Last edited:

Tropheus

n00b
Joined
Jun 8, 2011
Messages
18
5 Disk 4k Drive Raidz

My problem is that I already have the 4k drives otherwise I would have avoided them....

I am not concerned with the suggested 20% performance loss, my concerns are with the data if they are not aligned correctly on format. I have 6 WD 2tb hard drives and plan to setup a 5 disk raidz with the remaining disk set as a basic drive.

Gea thank you for your advise, I am thinking on the lines now of forgetting the align =12 patch and see what happens once I create the pool in just Napp-it. 3 of the drives are already formatted as a 3 disk raidz can I destroy this pool and recreate the 5 disk raidz in napp-it or do I have to SSH to wipe the existing drives
 

_Gea

2[H]4U
Joined
Dec 5, 2010
Messages
4,065
just destroy the pool and recreate a new one.
and don't care about data security. its only a slight performance issue.
(beside other problems like the firmware bug of Samsung F4's but thats a completely different issue)

Gea
 

Astronot

Weaksauce
Joined
Feb 1, 2008
Messages
83
It seems like no matter what i do it fails.

Now time sync works but thats it.

Try to join domain :

step 1: Timesync with ad-server adds: 15 Jun 07:56:28 ntpdate[2966]: adjust time server 192.168.1.19 offset -0.094527 sec
step 2: set dns entry in /etc/resolv.conf:
step 3: set /etc/krb5/krb5.conf:
step 4: try to join via: smbadm join -u administrator freak - please wait ....



Joining freak ... this may take a minute ...


failed to find any domain controllers for freak

this is the dns it makes:
search freak
domain freak
nameserver adds
(I tried changing freak to freak.local i tried changing adds to to IP)

Here is the krb5
[libdefaults]
default_realm = FREAK

[realms]
FREAK = {
kdc = adds
admin_server = adds
kpasswd_server = adds
kpasswd_protocol = SET_CHANGE
}

[domain_realm]
.freak = FREAK

(I tried changing adds to ip, freak, freak.local)

When i do it manually in terminal i always get failed to find any domain controllers for .... (i tried both freak and freak.local)

Did you do tihs part?

Now, set up /etc/nsswitch.conf so that hosts are resolved via DNS. You can modify you existing/etc/nsswitch.conf file or just copy /etc/nsswitch.dns to /etc/nsswitch.conf.

Also, if your domain controller is named adds, you should enter as adds.freak.local
 

Freak1

Limp Gawd
Joined
Sep 9, 2009
Messages
191
Did you do tihs part?



Also, if your domain controller is named adds, you should enter as adds.freak.local


Yes. I did all the steps.

I also did the nslookup for adds.freak.local works fine. And the ntpdate also gets a new time.

I even made a new VM with a new OpenIndiana and Napp-it, but its still the same.
 

axan

[H]ard|Gawd
Joined
Nov 5, 2005
Messages
1,935
is this 2008 Active Directory? if so you need to run this command before joining domain.

Code:
sharectl set -p lmauth_level=2 smb
 

s0rce

Limp Gawd
Joined
Jan 17, 2011
Messages
495
My boot drive on my little solaris 11 express server just died, its not recognized and the server won't boot. I have a recent snapshot that I sent to my data pool (as a file) about a week ago. I was wondering how I go about restoring the boot drive from this snapshot? I have another of the same drive that I can use.

Should I boot off the solaris cd, import my data pool and then can I zfs send/receive the snapshot to the empty drive?

Any help would be great!!

Thanks
 

s0rce

Limp Gawd
Joined
Jan 17, 2011
Messages
495
I think that might work, don't forget to set the boot blocks too.

sorry, I'm a bit new to all the ZFS stuff, how do I do that?

I'm not sure I have a full recursive snapshot (oops)...

edit: I'm reading and learning!
 
Last edited:

Freak1

Limp Gawd
Joined
Sep 9, 2009
Messages
191
is this 2008 Active Directory? if so you need to run this command before joining domain.

Code:
sharectl set -p lmauth_level=2 smb

You just saved my life :D

IT WORKS wrote that line then joined again "successfully" right away.

Thank you very much. Had i just known that a few days ago, it would have saved my some hours.
 

Freak1

Limp Gawd
Joined
Sep 9, 2009
Messages
191
I just popped one of my new 3TB ultrastar Hitachi in. But it only shows 2.2GB i know that that is a limit for some systems. But what is keeping OI/Napp-it from seeing the whole? I'm on a old LSI controller just for testing until my new LSI controller arrives. But i don't know if that is the culprit?

I was hoping to make a temporary RAIDz of these 3TB drives, to copy from my data from my raid60 setup. Before i get the new controller.

The setup if that matters
1x SuperMicro SC846A-R1200B (Changing backplane to Expander in the future)
1x SuperMicro 847E16-RJBOD1
1x MBD-X8DAH+-F -B
2x Intel Xeon E5606, 2.13 GHz - Quad Core/1066/8 MB
2x Heatsink SNK-P0038P (Rev. A & B) 2U+ DP Server
6x Kingston DDR3 ECC Reg, 1333 MHz, DR x4, 4 GB (24 GB)
1x Kingston USB Stick 8GB (For ESXi)
2x Intel 510 Series SSD, 120 GB, 450/210 MB/sec (RAID 1 For OI/Napp-it + 2008r2 AD)
1x Adaptec RAID controller for SSDs
2x SAS EL2/EL1 Cascading Cable (External), 68cm CBL-0166L
1x LSI SAS9201 PCIE 16 PORT INT (Waiting for this) An older is LSI is used now
1x 8GB Kingston USB stick for ESXi
 

jwinsor566

n00b
Joined
Feb 26, 2011
Messages
13
What parameters should I run with the Bonnie++ Benchmark? I am using Nexenta and have 12 GB RAM 10 (5x mirrored) Segate Constellation 7200 SAS 1 TB drives. I am only seeing ~ 60 MB/s right now ?? Is this normal I see a post in this thread where someone has 4 Spinpoint SATA f3's getting 160 MB/s what am I missing here??
 

s0rce

Limp Gawd
Joined
Jan 17, 2011
Messages
495
So my boot ssd just died and I've got a spare identical drive so I'm trying to install it and then load up a recent snapshot I backed up.

I've been following: http://www.solarisinternals.com/wik...ide#Replacing.2FRelabeling_the_Root_Pool_Disk

To recreate the drive and then: http://www.solarisinternals.com/wik...Guide#Complete_Solaris_ZFS_Root_Pool_Recovery

to recover from the snapshot.

I'm able to fdisk and partition the drive and then create the rpool but I can't seem to install grub

I ran:
raw device must be a root slice
installgrub installgrub /boot/grub/stage1 /boot/grub/stage2 /dev/rdsk/c9d0s0
and various permutations c9d0, etc

it always doesn't work and says something about raw device must be a root slice or a few. I'm going to try again and post the exact error messages. Any help would be great!!!

Thanks (please save my little server!)

edit:

for c9d0 it says

open: no such file or directory
unable to gather device information

for c9d0s0

cannot open/stat device
unable to gather device information

if I create the zpool before doing installgrub then it says invalid vdev specification
the following errors must be manually repaird, /dev/rsk/c9d0s0 is part of an active ZFS pool
 
Last edited:
Joined
May 27, 2011
Messages
34
My "screenshot"

ausset@asaheim:~$ zpool status
pool: rpool
state: ONLINE
scan: resilvered 9,82G in 0h2m with 0 errors on Fri Jun 17 22:40:38 2011
config:

NAME STATE READ WRITE CKSUM
rpool ONLINE 0 0 0
mirror-0 ONLINE 0 0 0
c2t0d0s0 ONLINE 0 0 0
c2t1d0s0 ONLINE 0 0 0

errors: No known data errors

pool: tank
state: ONLINE
scan: none requested
config:

NAME STATE READ WRITE CKSUM
tank ONLINE 0 0 0
raidz2-0 ONLINE 0 0 0
c1t0d0p0 ONLINE 0 0 0
c1t1d0p0 ONLINE 0 0 0
c1t2d0s2 ONLINE 0 0 0
c1t3d0p0 ONLINE 0 0 0
c1t4d0p0 ONLINE 0 0 0
c1t5d0p0 ONLINE 0 0 0

errors: No known data errors
 
Joined
May 27, 2011
Messages
34
So my boot ssd just died and I've got a spare identical drive so I'm trying to install it and then load up a recent snapshot I backed up.

I've been following: http://www.solarisinternals.com/wik...ide#Replacing.2FRelabeling_the_Root_Pool_Disk

To recreate the drive and then: http://www.solarisinternals.com/wik...Guide#Complete_Solaris_ZFS_Root_Pool_Recovery

to recover from the snapshot.

I'm able to fdisk and partition the drive and then create the rpool but I can't seem to install grub

I ran:
raw device must be a root slice
installgrub installgrub /boot/grub/stage1 /boot/grub/stage2 /dev/rdsk/c9d0s0
and various permutations c9d0, etc

it always doesn't work and says something about raw device must be a root slice or a few. I'm going to try again and post the exact error messages. Any help would be great!!!

Thanks (please save my little server!)

edit:

for c9d0 it says

open: no such file or directory
unable to gather device information

for c9d0s0

cannot open/stat device
unable to gather device information

if I create the zpool before doing installgrub then it says invalid vdev specification
the following errors must be manually repaird, /dev/rsk/c9d0s0 is part of an active ZFS pool

Use the commands as root

Do a " su " before doing anything. It worked for me when I was trying to mirror my root disk.
 
Joined
May 27, 2011
Messages
34
Gea, I used wget -O - www.napp-it.org / amp perl to install apache. But there is one problem :

apache was installed in /etc/apache2/2.2 instead of /etc/apache2/

How can I repair that?

I am trying to install VirtualBox ( done ) but I need help configuring Apache to install phpVirtualBox
 
Last edited:

Freak1

Limp Gawd
Joined
Sep 9, 2009
Messages
191
Am I'm right when i presume that everyone have access to the VM folder on napp-it if i do it like the all-in-one pdf? So anyone can go in and delete some files for the VM? I hope there is another way to do that, but is there?
 

_Gea

2[H]4U
Joined
Dec 5, 2010
Messages
4,065
Gea, I used wget -O - www.napp-it.org / amp perl to install apache. But there is one problem :

apache was installed in /etc/apache2/2.2 instead of /etc/apache2/

How can I repair that?

I am trying to install VirtualBox ( done ) but I need help configuring Apache to install phpVirtualBox


the napp-it installer is using the default Solaris packages with their default values. I don't know if you can modify the
installer. You may use links or you may move the files but then you have to check the service files and all apache files.

i would say it may be easier to keep Apache where it is in Solaris and try to adjust the things you like to run on Apache.


Gea
 

_Gea

2[H]4U
Joined
Dec 5, 2010
Messages
4,065
Am I'm right when i presume that everyone have access to the VM folder on napp-it if i do it like the all-in-one pdf? So anyone can go in and delete some files for the VM? I hope there is another way to do that, but is there?

You have to do it the same way like you would do with ESXi and a separate SAN.
The usual way is to define a secure SAN/ management-network and insecure LAN or WAN networks.

Only from this secure network you can allow access to your NFS/ ESXi configuration.
To do this, you usually have a physical switch, configured with vlans. One of the tagged
ports with all vlans are connected to the ESXi virtual switch.

You can then define virtual nics and connect them to the needed networks.
Other way is to use firewalling but thats not suggested.

But one thing is correct.
If you are in an unsecure network, you should not use an all-in-one as a ESXi/SAN and
as a common smb-filer at the same time unless you use extra security things like firewalling, zoning
or a second virtual nas-server with its own controller/disks.

Gea
 
Joined
May 27, 2011
Messages
34
the napp-it installer is using the default Solaris packages with their default values. I don't know if you can modify the
installer. You may use links or you may move the files but then you have to check the service files and all apache files.

i would say it may be easier to keep Apache where it is in Solaris and try to adjust the things you like to run on Apache.


Gea

Ok. But napp-it seems to be expecting Apache2 in /etc/apache2. Which is not the case
 

Zerosum

Limp Gawd
Joined
Feb 9, 2010
Messages
216
Would someone explain some quick zfs permissions for me?

Right now I'm on OI with a zfs filesystem and napp-it. I have smb shares working fine.

I create files/directories on my windows samba mapped directories. I can delete and overwrite those as expected. But files and directories created on the machine I cannot. I had to transfer and organize alot of other files and doing it via samba was slow. So I ended up doing it via shell.

How can I give these other files and directories the same type of permissions? I'm not worried about security as this is my own personal zfs server at home. I am the only user so I just log in and mount the same directories on all my other pcs using a single samba login. (Desktop, HTPC, Encoder, etc)

I'm not familiar with solaries/zfs ACLs. Basically I just want it world read/write/deleteable. I made sure the files are owned by the actual filesystem are owned by samba:staff (all the other ones are) and just chmod 777 everything.
 

_Gea

2[H]4U
Joined
Dec 5, 2010
Messages
4,065
Would someone explain some quick zfs permissions for me?

Right now I'm on OI with a zfs filesystem and napp-it. I have smb shares working fine.

I create files/directories on my windows samba mapped directories. I can delete and overwrite those as expected. But files and directories created on the machine I cannot. I had to transfer and organize alot of other files and doing it via samba was slow. So I ended up doing it via shell.

How can I give these other files and directories the same type of permissions? I'm not worried about security as this is my own personal zfs server at home. I am the only user so I just log in and mount the same directories on all my other pcs using a single samba login. (Desktop, HTPC, Encoder, etc)

I'm not familiar with solaries/zfs ACLs. Basically I just want it world read/write/deleteable. I made sure the files are owned by the actual filesystem are owned by samba:staff (all the other ones are) and just chmod 777 everything.

Kernel based SMB server (its SMB not Samba, Samba is an alternative SMB Server) is intended to be as
Windows compatible as possible for a unix machine. So its using ACL only, just like Windows.

What you can do:
set aclinherit on a shared ZFS folder to on, so ACL's are inherited. (=default, if you create with napp-it)
set ACL everyone@ to full + activate dir and file inherit of everyone@ ACL. (you can use the new acl-extension in napp-it to do this)

if you have problems with already created files, you may do a recursive chmod to 777
per default, new files created with CLI are root and owner editable only.

if you have done it as root, a mapping winuser:* unixuser:root may help also
(every Windows user has root equivalent file permissions)


Gea
 
Last edited:

WiseOldOne

n00b
Joined
Oct 28, 2010
Messages
11
Has anyone tried SAS multipath with an LSI9211-8i and a solaris flavoured os?

I'd really like to to this with IR firmware which passes the disks through, since this gives me ses, which is lacking in IT firmware.

Single controller gives 4 paths, and dual gives 8, but with oi147, the multipath is somewhat confused.
stmsboot shows a controller and four ports, rather than two controllers.


Chassis is SM847E26 with sas disks (just two for testing, but will use Seagate 2Tb SAS when it is working)
 
Joined
Dec 30, 2010
Messages
43
i'm trying to give a user access with sftp. how can i prevent this user from browsing my server? i set homefolder to /tank/ . add him to new group sftp. and add

Subsystem sftp internal-sftp

Match Group sftp

ChrootDirectory %h

ForceCommand internal-sftp

AllowTcpForwarding no

to ssh config to Chroot users with OpenSSH. with this config edit user cant login at all
 
Joined
May 27, 2011
Messages
34
While I was able to NFS into my server just after installing it I am now unable to access it with NFS with my Mac and Ubuntu machines.

I've got these messages : RPC not available on server, or sometime nfs.mount : server denied permission.

Also I did managed to install phpvirtualbox but I now have a question :

How do I autostart Virtualbox and it's VM ( and others ) in Openindiana ?
Also any idea on my disk naming "incoherence"?

How should I tune my system considering I have 16 Gb RAM and I run 2 VM with 6 Gb allocated in total ( 10 Gb remaining ) ?

Thanks for this great project !
 
Last edited:

El_Kurgan

n00b
Joined
Dec 8, 2010
Messages
20
Hi. I'm trying to mount my cifs shares from mi solaris 11 in nexenta but I'm getting this error:

mount_smbfs: service "svc:/network/smb/client:default" not enabled.


root@backup:~# svcs -xv
svc:/system/idmap:default (Native Identity Mapping Service)
State: disabled since Mon Jun 20 20:45:13 2011
Reason: Disabled by an administrator.
See: http://sun.com/msg/SMF-8000-05
See: man -M /usr/share/man -s 1M idmapd
See: man -M /usr/share/man -s 1M idmap
Impact: 2 dependent services are not running:
svc:/network/smb/client:default
svc:/network/smb/server:default

I've enabled it manually and using napp-it interface. Does anyone knows what happens??

thanks
 

ochadd

[H]ard|Gawd
Joined
May 9, 2008
Messages
1,101
Thanks for the post. NexentaCore install worked out after figuring out the following to make DHCP work:
svcadm enable nwam

Trying it out in virtualbox.
 
Joined
May 27, 2011
Messages
34
My server was a little noisy so I :

Set cpupm to " enable"
Set cpu threshold to "1s"
Set harddrive threshold for ALL drives EXCEPT bootdrives to 30 min

Disabled Solaris Fault Management.

Does it seems good? Is Solaris Fault Management that useful?

Thanks

Also :

I took a look at my logs, and I have HUNDREDS of :
" [ID 589989 daemon.error] Could not find DNS entry for tcp "
What can i do to have a working NFS?
 
Last edited:

cbloss

Limp Gawd
Joined
Feb 1, 2001
Messages
410
FYI I had to make symbolic links for the napp-it interface to see my disks because /dev/rdsk/c7t#d0 didn't exist.

ln -s /dev/rdsk/c7t0d0p0 /dev/rdsk/c7t0d0

I'm using Solaris 11 Express. Also, I just installed the newest version of smartmontools (5.41) and it seems to work without issues.
 
Top