VMware ESXi

VMware cluster and croit work together very well. It is possible to store VMs on Ceph RBD via the iSCSI protocol. With the help of our iSCSI service gateways we achieve a HA Multipath iSCSI setup, which can keep the VMs running even if individual nodes or network segments fail.

#Setting up iSCSI

#Add Group

  • Select the newly created iSCSI service
  • Press edit in the action bar at the bottom of your window
  • Press + Add Client in the action bar at the bottom of your window

#Add Image to group

  • Navigate into the group register and select the newly created group
  • Press + Add Image in the action bar at the bottom of your window

Repeat that step until you have enough images configured.

Please note that for Multipath at least the number of images must be available as paths want to use them.

Vmware iscsi iqn

# Add Client using the VMware IQN

  • Open your VMware Webinterface in another browser tab
  • Navigate into Storage and select Software-iSCSI
  • A configuration dialog will show up and provide you a Name and Alias. Copy it!
  • Navigate back to your croit iSCSI configuration dialog
  • Paste the IQN string from the VMware Interface into the Dialog
  • Add a Username of your choice
  • Add a Passwort of your choice, or generate one
  • Select the previously created iSCSI group

# Save iSCSI configuration in VMware

  • Navigate to your VMware Webinterface
  • Configure the username and password
  • Expand the CHAP authentication section, select Do not use CHAP unless required by target
    The Mutual CHAP authentication section should have Do not use CHAP selected
  • Save the configuration and verify that everything is working as expected
Vmware iscsi dialog

#Set iSCSI recovery timeout in VMware

Expand Advanced settings and set the “RecoveryTimeout” to 25.

Vmware iscsi recovery timeout

VMWare tends to have pathing issues with ceph-iscsi, causing the gateways to crash. Set this:

esxcli storage nmp satp rule add -V LIO-ORG -M "TCMU device" -P VMW_PSP_MRU -s VMW_SATP_ALUA
esxcli storage core claimrule load
And reboot your hypervisor afterwards.