forcefully remove a node from virtual Isilon

83 views
Skip to first unread message

Robert Soeting

unread,
Mar 19, 2020, 11:41:30 AM3/19/20
to Isilon Technical User Group
Does anyone know how to forcefully remove a virtual (test Isilon) from a cluster
Or how to add ip's to a internal subnet as our test Isilon complains about "Failover is not supported".
I am trying to fix a broken node by doing a smartfail and then add a new node. Someone tried replacing some binaries on the broken node (don't ask me why).
To make it more clear, I am not talking about a EDGE Isilon. These are the OVA files you can download from the support site.
Reason why I not throw away the whole cluster and start over is because we have customized a lot which would take very long to replace.

Any help is welcome.

Adam Fox

unread,
Mar 19, 2020, 11:50:52 AM3/19/20
to isilon-u...@googlegroups.com
I haven’t tried it in a long time, but I would think that ‘isi devices node smatrfail’ would work to remove a node just like on a regular cluster.
To update the internal interfaces, you should be able to use the iprange command from inside isi config.

Hope this helps

— Adam Fox


--
You received this message because you are subscribed to the Google Groups "Isilon Technical User Group" group.
To unsubscribe from this group and stop receiving emails from it, send an email to isilon-user-gr...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/isilon-user-group/b71c3c03-bf3f-4c99-b1f3-b0ec93c337d3%40googlegroups.com.

Robert Soeting

unread,
Mar 19, 2020, 12:05:39 PM3/19/20
to Isilon Technical User Group
Hi Adam,

Thanks, completely forgot about the "isi config" part. Iprange is extented so we can add additional nodes now. 
Now finding a way to get rid of the pesty broken node. Smartfail is not working we tried that multiple times. Node is now in a bootloop and cluster is just saying D-S- (Down Smartfailed). But does not evict the node from the cluster.

Thanks for the help so far. 
Robert

On Thursday, March 19, 2020 at 4:50:52 PM UTC+1, Adam Fox wrote:
I haven’t tried it in a long time, but I would think that ‘isi devices node smatrfail’ would work to remove a node just like on a regular cluster.
To update the internal interfaces, you should be able to use the iprange command from inside isi config.

Hope this helps

— Adam Fox

On Mar 19, 2020, at 11:41 AM, Robert Soeting <robert...@gmail.com> wrote:

Does anyone know how to forcefully remove a virtual (test Isilon) from a cluster
Or how to add ip's to a internal subnet as our test Isilon complains about "Failover is not supported".
I am trying to fix a broken node by doing a smartfail and then add a new node. Someone tried replacing some binaries on the broken node (don't ask me why).
To make it more clear, I am not talking about a EDGE Isilon. These are the OVA files you can download from the support site.
Reason why I not throw away the whole cluster and start over is because we have customized a lot which would take very long to replace.

Any help is welcome.

--
You received this message because you are subscribed to the Google Groups "Isilon Technical User Group" group.
To unsubscribe from this group and stop receiving emails from it, send an email to isilon-u...@googlegroups.com.

Erik Weiman

unread,
Mar 19, 2020, 12:32:42 PM3/19/20
to isilon-u...@googlegroups.com
keep the bad VM node powered off and try running flex in that state. 
As long as you have quorum and aren't over your protection level it should generally just work. 

To unsubscribe from this group and stop receiving emails from it, send an email to isilon-user-gr...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/isilon-user-group/d8960a31-f232-4597-b736-7df12086d08d%40googlegroups.com.

Peter Serocka

unread,
Mar 19, 2020, 12:41:19 PM3/19/20
to isilon-u...@googlegroups.com
As a general protection before further experiments,
shutdown the whole cluster and take a snapshot for each node VM as such,
i.e. a VMware snapshot. So you can go back to this point in time
later if needed. (Slight advantage for virtual vs physical Isilons :)

Next, start the healthy nodes and do a stopfail instead of a 
smartfail on the bad one.
Add a fresh node and let the cluster run flexprotect.

hth
-- Peter


Reply all
Reply to author
Forward
0 new messages