ISCSI getting disconnected vsphere 7.0 (dead)

Software-based VM-centric and flash-friendly VM storage + free version

Moderators: anton (staff), art (staff), Max (staff), Anatoly (staff)

Post Reply
MohamadB
Posts: 4
Joined: Thu Jun 04, 2020 12:56 pm

Thu Jun 04, 2020 3:32 pm

Hi all
I have a problem with vSphere 7.0
configuration as below
x2 node ESXi same config
SSD = datastore 2tb
HDD = datastore 4tb
128gb ram
x6 nic 10G
StarWind over windows core 2016
32gb RAM
os = 15gb SSD datastore Thick Provision Lazy Zeroed VMware Paravirtual
SSD = image ha 1.8gb SSD datastore Thick Provision Lazy Zeroed VMware Paravirtual
HDD = image ha 4tb HDD datastore Thick Provision Lazy Zeroed VMware Paravirtual
nic1 VM Management E1000e 1500 1.1.1.1
nic2 heartbeat VMXNET 3 jumbo 9000 2.2.2.1 BTB
nic3 sync1 VMXNET 3 jumbo 9000 3.3.3.1 BTB
nic4 sync2 VMXNET 3 jumbo 9000 4.4.4.1 BTB
nic5 FT jumbo 9000 BTB
nic6 VMotion jumbo 9000 BTB

all ping ok

STORAGE CONFIGURATION both node have the same config
ISCSi added ok
Dynamic Discovery
2.2.2.1 node(2) 2 2.2.2 heartbeat interface
2.2.2.2 node(2) 2.2.2.1 heartbeat interface
Network Port Binding
nic interface VKernel heartbeat = last active
round-robin path policy.
I see 4 targets which mean configuration ok :
all Active I/0

I created the first datastore from the first HA image
and the second datastore from the second HA image

all good all are synced no problem

I created new Virtual Machine with 1 disk 100GB Thick Provision Lazy Zeroed I select datastore for the VMDKS to be stored in
first HA image all good no problem I installed windows server 2016 all perfect

now I want to enable fault tolerance for this new virtual machine I select enable
option ask me where and which HA datastore will use I select the second datastore no problem all good
when I start the virtual machine first VM run and fault tolerance start to copy the virtual machine to the second HA image after 2 min ISCSi target for the second HA image get
disconnected and dead, what I did
I changed the SSD
I changed the HDD
I deleted all the targets recreate all from scratch more the 3 times same problem without fault tolerance which mean without using the second HA image at the same time first HA being used
nothing happens

I need to know is there any mistake with my configuration

note: I created 4TB HA image using Think-provisioned LSWB (with log) block size 512 Mode (WT) ram cache 4GB (Do I need more RAM )
1.8 TB HA image I used 2GB for ram cache (Do I need more RAM )
log saved on SSD for both HA image 2TB size (Do I need more space) not the same SSD that been used for the SSD HA image

Thanks, ALL

Regards,
Michael (staff)
Staff
Posts: 317
Joined: Thu Jul 21, 2016 10:16 am

Mon Jun 08, 2020 4:30 am

Hello MohamadB,
Thank you for your interest in StarWind products.
To configure the setup, please follow this guide: https://www.starwindsoftware.com/resour ... phere-6-5/
Please note the following:
- SSD and HDD RAID settings: https://knowledgebase.starwindsoftware. ... ssd-disks/
- in the guide, it's recommended using thick-provisioned eager zeroed disks;
- check the number of vcpu you have assigned to VM with StarWind VSAN;
- reserve compute resources for the VM with StarWind VSAN;
- you have configured public IP address, use private instead;
- remove binding for storage;
- LSWB is an experimental feature - use simple image disks instead;
- use L1 cache only for HDD devices, SSD devices can be configured without cache.
Please let us know the result.
Post Reply