Unsolved
4 Posts
0
1619
Isilon Simulator 8.2.2.0 : Error joining cluster: no response from node: Operation timed out
Hi, I am setting up an Isilon 8.2.2 simulated cluster on VMware Workstation Pro 15.5 on Centos 8.
The first node install seemed to go okay although there are warnings about DNS entries missing.
I configured the int-a addresses using the HostOnly network, 172.16.221.100-110 and the ext-a addresses using the bridged network, 10.0.0.100-110.
From the host computer, I can connect to both 172.16.221.100 and 10.0.0.100 to login to the 1st node.
But when I try to install the 2nd node, at the point of joining the cluster, the cluster is seen with status available.. But not joined, the error is Error joining cluster: no response from node: Operation timed out
I assume it is using something similar to DHCP to get IP address information from the 1st node, but I don't know how to figure out why this is failing, I have done this before without seeing this error.
Thanks
PunctuatedEquilbrium
4 Posts
0
June 30th, 2020 14:00
Update, I thought the issue might be tied to the configuration of the vmnet interfaces in VMware Workstation but I have tried different configurations without success.
Right now I have:
vmnet0 : Host-Only -> em0 (int-a)
vmnet1 : Bridged -> em1 (ext-a)
As far as I can tell the networking on the first node is all working fine. I tried doing packet captures in VMware Workstation using vmnet-sniffer and in the Isilon Admin GUI using NetGather. It looks like the 2nd node gets the information about the existence of the first node using LLDP. And when it tries to connect to the first node it sends UDP broadcasts on port 5041 which are received by the first node but no response is sent.
Thanks
DELL-Sam L
Moderator
Moderator
•
6.9K Posts
0
July 1st, 2020 10:00
Hello PunctuatedEquilbrium,
I would double check to make sure that you are installing your cluster as stated in the install guide. https://dell.to/2YOX9Q3
Here is also a link to a KB that could also help. https://dell.to/2YQ4FKH
Please let us know if you have any other questions.
Phil.Lam
1 Rookie
1 Rookie
•
573 Posts
0
July 2nd, 2020 09:00
@PunctuatedEquilbrium
Go into node 2 console to join the cluster, it not automatic. Or from node 1 do these commands:
isi devices node list
isi devices node add --serial-number=
PunctuatedEquilbrium
4 Posts
0
July 2nd, 2020 09:00
I was able to login to WebUI on Node 1 and add the additional nodes through Cluster Management->Hardware Configuration
The doc is not very detailed about the VM network configuration, I ended up using 2 'custom':
em0 -> /dev/vmnet0 (host-only) for int-a
em1 -> /dev/vmnet1 (bridged) ext-a
Thanks
Geoff
tenortim
36 Posts
0
July 23rd, 2020 12:00
Hi Geoff,
From the cluster side, em0 is int-a (first back end network), em1 is ext-1 (front end network), and em2 is int-b (second back end network).
If you're trying to map the OVA names to actual interfaces networks (e.g. in ESXi), the mapping is
Even when deploying in workstation, I'd suggest creating two private network switches for the back end bridging the front end (the network labeled "nat") so that you can test client access to the simulator.