Netapp Simulator Change Serial Number
Ania Cozzolino
I'm building the second node as per the install instructions but am getting a system id mismatch error after setting the env variables SYS_SERIAL_NUM and bootarg.nvram.sysid. Any suggestions? This is the 8.2.1RC1 version of the simulator in cluster-ontap mode.
I didn't see a cf vmdk, but I did replace the DataONTAP-var.vmdk and DataONTAP-sim.vmdk files with original ones from the tgz package. I had also found and edited a file called env within the cfcard\env directory of the virtual machine folder. Replacing just the -var.vmdk didn't work, but after editing the env file with the updated serial info I used in the instructions and replacing the -sim.vmdk file, I got past the error that prevented the simulator from booting. I don't know if editing the env file directly had anything to do with this, but this file still contained the original SYS_SERIAL_NUM and bootarg.nvram.sysid.
Glad you got it working. The env file is a remnant of the build process, and not actually used by the simulator. A copy of it lives on the boot device, DataONTAP.vmdk - historically a compactflash card, and represents the initial boot environment for the simulator. If you are going to make changes, such as modifying the serial, switching modes or adjusting the simdisk inventory, its best to catch the very first boot. During the first boot, the nvram is discarded, /var is populated, and /sim is populated with simdisks. If you miss the first window and have problems its easier to start over with clean files than it is to reset and re-arm those initialization processes.
Is there a (documented) way to get several simulators into Operations Manager? It seems that due to the fact that mny sims all have the same serialno/SystemID so OM seems to think they all the same machines. I am trying to get a Protection Manager Demo running and need at least two 8.0(.1) sims for that...
I have found that at bootup the Sim looks in /sim/dev/ for a file called ,serialno (just like the 7.x Sims), but I cannot access that file, as the systemshell demo user does not have right to create that file...
The one thing I'd add is that if you make that serial number change in the SIMLOADER before first setting up the simulator, then it won't be necessary to go through the maintenance mode process to reassign the disks. The initial setup will use the serial number that's defined at the time.
Just a quick addendum: I've recently found out that there's another boot variable that takes precedent in setting the UID's. If you're looking to avoid UID conflicts, the best thing to do is set bootarg.nvram.sysid and SYS_SERIAL_NUM to the same value, and do this prior to first boot and initializing the disks. That prevents having to go through maintenance mode to reassign the disks. Other UID's may also be fixed and harder to change after the first complete boot.
I tried this method on a fresh install of the sim, I drop to SIMLOADER and enter in the "set" commands listed, boot, hit ctrl-C to enter the boot menu, and select "4" to initialize the disks and wipe config. Then the VM dumps core and reboots without doing the config wipe / disk init.
Are you using one of the original 8.0.0 simulators? That method may not work on the older vsims. If this is happening with the DOT 8.0.1 vsim, then the only thing that comes to mind is to make sure that the above steps are performed as the first thing after the simulator is extracted from the zip file and powered on.
Yes, seems that happens in some DOT8.x simulator versions. In that case follow these steps to change the serial number (you'll need a fresh simulator to do this, can't use one that already ended up in the panic situation).
I downloaded vSIM 8.1, installed it on ESX4i, at first boot stopped in SIMLOADER and changed both SYS_SERIAL_NUM and bootarg.nvram.sysid (just one number down from original). I then continued with SIMLOADER> boot and stopped in special boot menu, selecting option 4 for a fresh install.
I installed two ONTAP 9.5P6 simulator nodes and changed the SN ofone node following the installation guide. I configured two single node clusters based on these two nodes. Then I found the cluster SN of both clusters is 1-80-000011.
NetApp ONTAP disk management using the simulator is a great opportunity to learn about the NetApp offering. In particular for those who run a homelab it is a nice kit to add and understand more about the configuration. When installing the NetApp ONTAP simulator for the first time it deploys with a standard configuration. It consists in 28 emulated disks over 2 shelves. In reality the NetApp ONTAP simulator allows up to 4 disk shelves with disks up to 9 GB each. But what expanding or changing the default configuration?
This is what this article is about: deleting the built-in disk configuration and create new ones from scratch. It goes without saying a backup of the data sitting on the ONTAP simulator should be protected prior to follow any further step. Literally the existing disks (emulated drives in the ONTAP simulator) will be wiped and recreated again.
This article explores the option to delete the existing disk configuration, remove the emulated disks and create new ones based on specific templates. In particular 3 tiers for Capacity, Performance and Ultra-Performance. These come handy in homelab setup when simulating different tiers of storage differentiating between standard and critical applications. More on this later on a dedicated article.
The process might look quite lengthy. In reality is very quick. Once acknowledged the commands the procedure is straight forward. This means the NetApp ONTAP simulator could be configured up to half terabyte of storage. That is 9 GB x 56 disk slots over 4 disk shelves. A whopping 504 GB. Not bad for homelab! In reality the NetApp ONTAP simulator is limited to 220 GB by default. In order to extend this to 500 GB there is another hack. But this will be covered in a second part with detailed step by step. 220 GB for a homelab are enough to cover many scenario.
One of the first things to change the NetApp ONTAP disk configuration is to connect to the ONTAP appliance using an SSH connection. This is enabled by default. Otherwise this can be configured in the default OnCommand System Manager as part of the protocols allowed for the users.
In this case the Putty utility is used for the SSH connection to the ONTAP appliance. Once connected and authenticated a simple Disk show will reveal the main information about the disk details including the Node owners in the cluster.
Next is enable the Advanced mode with the set -privilege advanced and confirm with a yes. This step might be required from NetApp ONTAP 9.4 and later. It is also worth noting the * symbol on the shell with next commands in this context.
This will now show all the NetApp ONTAP disk types that can be generated. The very first number on the left column represents the ID for such disks types. For Reference a simple output for all disk types can be downloaded here.
At this point it is possible to start creating the disks based on the desired ID. To create the Ultra-Performance tier made of SSD disks the command would be similar to sudo vsim_makedisks -n 10 -t 35 -a 0
As expected these disks are not assigned to any Node owner yet. A quick test can be done by manually assigning these disk to the current node. This can be easily accomplished by running the disk assign command followed by the disk names.
Passionate about Virtualization, Storage, Data Availability and Software Defined Data Center technologies. The aim of Domalab.com is sharing with the Community the knowledge and experience gained with customers, industry leaders and like minded peers. Always open to constructive feedback and new challenges.
Hi Mike,
I was happy with the 220GB for my homelab and did not need 500+ for additional testing ? I found this video that goes with similar steps. Basically you need to detach the vmdk from the NetApp Sim, change/edit the vmdk to thin provisioning and re-attach again. This video covers the details =JWJuXS7WmAk. Hope this helps. Thanks, Michele
After changing the serial number control + C no longer works to enter the boot menu. I get the message: 'System IDs do not match'. I think this is because there is no real NVRAM to write the new system ID to. So I choose y to over ride the sys ID and get the following:
If you get that message then it has boot looped at least once before you made the changes in the loader. The simplest fix is delete it and deploy another one in its place. Open the VM console before you start it so you can catch it during the 10 second autoboot timer on the first boot.
In general the second simulator is configured like the first one. As a main difference you have to change the serial number and system id of the second VM and use the appropiate license numbers like already mentioned in Part 1.
After running the simulator for many years now I have experienced a wide range of scenarios that have required intervention. Many of the scenarios in this mega thread are based upon a mix of my own experiences, NetApp PS engineers knowledge and other community postings... oh and of course the instruction manuals! The goal of this thread is to bring it all together and show you the best practices which will in many cases prevent frustration further down the line. I hope you find this useful - I've tried to credit the sources as best I can remember.
I assume that you have already uploaded the relevant images to your datastores by extracting the .tar file from the support.netapp.com link and using the VMware datastore browser - yes its like watching paint dry and the "upload minutes left" may as well be a random number generator.
VMware ESX cannot find the virtual disk '/vmfs/volumes/DataONTAP-sim.vmdk' or simliar. Dont worry! This isn't a problem with the image, ESX doesnt load the multiextent modules required by default - it used to pre ESX 5.1 - grab yourself a copy or your favourite SSH tool and run the below:
7fc3f7cf58