I have written previously on deploying the ONTAP simulator on a vSphere host, and they seem to be some of my more popular posts. Now that the ONTAP 8.1 simulator has been released, I thought I would do an updated post that is a little more comprehensive than my previous posts.
In my most recent post on this topic, I ran through the configuration I use to optimize usable space and getting necessary license keys installed. I wanted to build on that for the 8.1 simulator, and specifically about getting multiple simulators to work with OnCommand. If you install multiple copies of the simulator and then try to add them into the NetApp Management Console, you will get an error similar to this:
Even though they are separate virtual machines and have different IP addresses and hostnames, their simulator system id’s are identical so the NMC thinks it’s a duplicate host you are trying to add. Luckily it is possible to change the serial number and system id so we can get multiple simulators added in and be able to utilize things like Protection Manager and Provisioning Manager. The easiest way to do this is to change these values before you run the setup so you don’t have to reassign the disks.
To begin, I still utilize VMware converter to bring it into my vSphere environment. There are other methods out there but this one has worked well for me. I won’t screenshot the entire process as it’s mostly just taking the defaults and deciding what name you want to use etc, but here is the summary screen:
Summary screen for VMware Converter
I change the disk type to be thin provisioned instead of thick, I change the NICs to be the relevant networks for my lab, and finally I didn’t set it to power on just because I wanted to make sure I boot it into maintenance mode to change the serial number as soon as it boots up. When it boots press a key other than enter to break the boot, then run the following commands to make this simulator unique (screenshot below):
set bootarg.nvram.sysid=1111111101 set SYS_SERIAL_NUM=1111111101 boot
Interrupt boot process to change serial number
These commands were documented in this post on the NetApp Communities, and I’ve followed the same pattern – 8 1′s with a unique two character string at the end that matches the hostname (e.g. STO-FAS1 is 01 STO-FAS2 is 02 etc)
After you hit enter to boot the simulator, you need to go into maintenance mode to wipe the config and setup the simulator, you need to hit control-C to bring up the boot menu and then option 4 for wiping the configuration:
Boot menu options
Once you run through the wizard, this is the config I apply (I usually use a text snippet to insert the following into my SSH session):
options security.passwd.rules.enable off snap reserve -A aggr0 0 snap sched -A aggr0 0 options autosupport.enable off aggr options aggr0 raidsize 28 disk assign all license add DZDACHD license add PZKEAZL yes license add NAZOMKC license add ANLEAZL license add BSLRLTG license add NQBYFJJ license add ELNRLTG license add MTVVGAF license add BQOEAZL license add RKBAFSN license add HNGEAZL license add BCJEAZL license add DFVXFJJ license add XJQIVFK license add DNDCBQH license add JQAACHD license add ZYICXLC license add PVOIVFK license add PDXMQMI license add RQAYBFE license add ZOFNMID license add ZOPRKAM license add RIQTKCL ndmpd on options nfs.export.auto-update off
I turn off the password rules since this is just my lab and I typically use a very easy password (yep, you probably already guessed it) for most of the lab stuff that can only be accessed from inside.
Next step is to add some more disks in the simulator, this info was found on this thread over on the NetApp Communities. Not all of the commands worked for me, in part 2 step 3 I wasn’t able to successfully enter those commands. It mentions that was a glitch in the way the program was complied so I’m just guessing that it may now be resolved and not necessary now as without it I still was able to add disks without issue.
priv set advanced useradmin diaguser unlock useradmin diaguser password
Enter a password to use for the diaguser (again, it’s my lab so I use a relaxed password)
systemshell
Login with the user diag and the password you just created
setenv PATH "${PATH}:/sim/bin"
cd /sim/dev
sudo makedisks.main -n 14 -t 23 -a 2
sudo makedisks.main -n 14 -t 23 -a 3
exit
useradmin diaguser lock
priv set admin
reboot
After the system comes back up, login and we can assign the new disks
disk assign all
If you run into an error where you get a bad disk label, that can be easily fixed by:
aggr status -f
This output will give you the list of the failed disks, make note of the disk ids – they should be similar to v6.32
priv set advanced disk unfail -s v6.32
At this point you should have a total of 56 disks on the simulator, with an aggregate raid group size of 28. I added 52 of the disks into my aggregate to leave 1 as a spare so I don’t constantly get errors about low spare count (you can disable the option to warn you about low spares but unfortunately that only works on systems with 16 disks or less).
aggr add aggr0 52
Now you should have the first simulator completed, I took a VM snapshot at this point to be able to revert to after I do my testing. Next up is configuring the second simulator, obviously the VMware Converter steps are basically identical except for the VM name so I won’t repeat that part. The only thing to be aware of is when you first power on the VM, press a key other than enter just like you did previously so we can set a different serial number:
Configuring unique serial number and system id on second simulator instance
Run through the rest of the configuration steps above and now you should be able to add both simulators into the NetApp Management Console:
Both simulators showing within the NetApp Management Console. Click for larger image.
Hopefully that helps you get started with the ONTAP 8.1 simulator, I’ll have some more posts coming up that will build off of this.
Popularity: 35% [?]
At a glance you can see todays PUE, as well as the months average among other things. It also shows the outside air info, being from Minnesota my first thought when I landed in Raleigh was how amazing the weather is – and this datacenter takes advantage of it. Since approximately 60% of the time the weather is 70 degrees or cooler, it has the ability to automatically pull in outside air as much as possible to help in cooling the datacenter. The average PUE for this datacenter is around 1.2 – one of the lowest in the industry, having a PUE of 1.2 gives an estimated savings of $7 million in operating expenses annually versus a PUE of 2.0.
And this picture is a close up of the gate we had to go through to get out of the lobby and into the building:
Once through the gate, we walked down a hallway and about halfway down you could see into their NOC(gotta say, the cubes are a little nicer than mine…):
Inside the datacenter there were a number of these(of note is the enclosed hot aisles, our tour guide also mentioned that they currently have the lowest PUE of any datacenter):
And the last picture showing the 3 different feeds, none of which are ever over 66% utilization:
My Popular Posts
