alekseylib
1 Copper

Re: VxRAIL Resetting Script

Jump to solution

It does not really matter where you add it, but this is one possibility...

C:\Python27\python.exe c:\reset\reset.pyc -d --destroyAllVMs --eraseVSANObjects -f <IPv4_node1> ... and so on....

0 Kudos
HyperConverge
1 Copper

Re: VxRAIL Resetting Script

Jump to solution

Hi, I will practice destroying VSAN once we get our dummy appliance.

Question:

If Mr. customer had an appliance which was wrongly reset without specifying the --eraseVSANobject switch.

Is there any means to identify which VSAN object is not being used so that it can be deleted to save space.

Thanks.

0 Kudos
alekseylib
1 Copper

Re: VxRAIL Resetting Script

Jump to solution

Is there any means to identify which VSAN object is not being used so that it can be deleted to save space.

Yes, but not to the mortal humans like you and I.

This requires a support call to VMW which can only be made by EMC support folks covering VXRAIL. VMW teams (vSAN specialists) have access to a special tool that can find stale / unused objects and delete them.

Technically, things like ISO images (that you mentioned in earlier comments) could be deleted via datastore browser if they show up there.

Please don't attempt to delete anything using cli commands from ESXi (i.e. rm). That's not supported and can cause other problems.

0 Kudos

Re: VxRAIL Resetting Script

Jump to solution

Hi there.

Just in case that anyone searching for a reset procedure for a VxRail Gen3 system finds that article here:
NEVER EVER use the reset script on a gen3 Node. Even if the support tells you so. If you do use the script an Gen3 hardware it actually starts disabling the VMs but fails in a certain step, leaving the nodes in an unusable state.

The reset procedure for Gen3 is integrated in the system itself. You need to connect via KVM directly to the node and during the boot process you see an additional boot option (called RASR boot) to select. Use that one to boot into some sort of service mode where one selection is resetting the node. If you try this on a script resetted node, you will see bad block errors preventing the reset function from running. If you do see this, just reboot the node once more, repeating the exact same steps like before, and everything will work in the second attempt. Don´t ask me why that actually happens, it did on ten nodes in a crashed by script cluster for me. Support told me do use the golden image and reimage the nodes what would have taken a lot more time to do.

I´m not sure wether or not all support staff is actually aware of that issue so this is a "just in case" information. 🙂

0 Kudos
alekseylib
1 Copper

Re: VxRAIL Resetting Script

Jump to solution

That's not supposed to happen. Feel free to ping me (privately) with the Dell EMC case number, so we can reach out to the support leadership team and RCA this.

Thanks.

0 Kudos

Re: VxRAIL Resetting Script

Jump to solution

Of course it´s not. 🙂

It happened February this year twice with the Pearl Script solution. It might have been changed by now so maybe it won´t happen again, I only wanted to be sure, that the correct reset procedure is used with Gen3 systems. The case is already "solved" by now with the described solution. But of course you can check the related SR. I´ll send you the number via PN.

Thanks for the prompt response.

Regards

Terence

0 Kudos
HyperConverge
1 Copper

Re: VxRAIL Resetting Script

Jump to solution

Hi,

I need to be able to have nodes reset but after reset, nodes need to be on the original IPs.

As example i have 4 nodes as follows:

N1:10.50.50.115

N2:10.50.50.116

N3:10.50.50.117

N4:10.50.50.118

After i had done the reset, all nodes goes to APIPA (169,254,x.x).

If there is DHCP on network it will get worse, i have to quickly go on each node and change ips for reset to continue.

Below is the script i run to reset a VXRAIL 4.0.301 Gen1 phoenix appliance.

C:\Python27\python.exe c:\reset4\reset\reset.pyc -d --destroyAllVMs --eraseVSANObjects -f 10.50.50.115 -a 10.50.50.116 -a 10.50.50.117 -a 10.50.50.118 --vlan 0 --initialIpAddress=10.50.50.204 --initialSubnet=255.0.0.0 --initialGateway=10.10.10.10

Where i need to add additional parameters in the command to have all nodes retain thier ips as 10.50.50.115 ect..

Thank you to help me

Regards

Wassim

0 Kudos
giadarola
2 Iron

Re: VxRAIL Resetting Script

Jump to solution

If you run the reset using the 169,254,x.x addresses instead of the post configuration IP addresses, then you will not have this issue and the reset will run thru to completion.


G

0 Kudos
HyperConverge
1 Copper

Re: VxRAIL Resetting Script

Jump to solution

Hi Gary-VxRail,

Thanks your reply. So i have no option to put IP back to 10.50.50.x after the script is running? . I will have to go to 169.254.x.x.

I know that on 169.254.x.x it works very well but what if customer wants to have the appliance be reset remotely outside its datacenter and VXrail appliance needs to be back on customer IPs so that they can access each esxi nodes.

Or by just accessing the VXRAIL manager custom IP i will be fine as it will be already set on custom nework via the --initialIPAddress parameter.

0 Kudos
giadarola
2 Iron

Re: VxRAIL Resetting Script

Jump to solution

GoodEMC,

When the VxRail appliance is being re-set, that sets it back to a factory fresh state. Once the reset is completed, then the first run configuration needs to be re-run to configure the VxRail cluster again.

Resets can be done remotely as long as you have a jump box that is multi-homes so that it can talk to the as-configured IP addresses as well as the 169.254.x.x self-assigned IP's.

We do this in our lab everyday.

If the ESX hosts have IP addresses hard set before first run, then the first run validation pre-checks will fail.

Remember that VxRail is not a DIY - Build your own. It is an engineers solution that has it's own set-up orchestration engine. That is the VxRail Manager UI.

HtH.

G

0 Kudos