Storage Boards Cleanup
To make it easier to find information about HPE Storage products and solutions, we are doing spring cleaning. This includes consolidation of some older boards, and a simpler structure that more accurately reflects how people use HPE Storage.
HPE StoreVirtual Storage / LeftHand
cancel
Showing results for 
Search instead for 
Did you mean: 

P4500 Multi-Site SAN iSCSI Initiator problems

Steve_Paine
Occasional Contributor

P4500 Multi-Site SAN iSCSI Initiator problems

We have a 6 node P4500 multi-site cluster connected to two ProCurve 5406zl switches at each site, these switches are for iSCSI only. We have 4 DL380G6 servers at each site running VMware ESX.. all iSCSI is running at 10Gb. ESXi 4.1 and SAN/iQ v9 all patched to the latest.

My problem is... we had a virtual machine running on a physical ESX host, we pull both connections to the storage expecting the VM to failover to another host, NO!! it just keeps running, all be it in a semi responsive state, we could ping it, and in some way still access it via a remote session, we could also issue commands although they did not respond back. Then, once the storage was reconnected the machine would continue to operate normally, the strange thing was, all the commands entered while the storage was unplugged then executed.. This is not what we would expect. How does a VM continue to operate with no disk? Is VMware caching the VM’s? is there a iSCSI time out issue? I thinks it’s an iSCSI time out issue with the VMware iSCSI initiator but I cant find any setting.

Can anyone help?
3 REPLIES
Amar_Joshi
Honored Contributor

Re: P4500 Multi-Site SAN iSCSI Initiator problems

Steve,
It will be helpful to know which hardware ESX is running on?

Usually, iSCSI initiator will timeout in less than 2 seconds if cable is plugged out. Surely this is an issue with network adapter (don't think that it's vSwitch or VMKernel setting). When you unplug the cable what's the status of network adapter in ESX? See it in GUI as well as with following commands:
# esxcfg-nics -l
# esxcfg-vswitch -l

Most likely ethernet adapter is not passing the failure status to ESX hence DRS/HA will not move the machine to another host.

Please also tell us if this is happening on all the hosts or just one?
Patrick Terlisten
Honored Contributor

Re: P4500 Multi-Site SAN iSCSI Initiator problems

Hello,

AFAIK a storage failover will not result in a failover of the VM. VMware HA will only act, if a physical host crashed. In you case you pulled both iSCSI uplinks from the host to the storage. This is a double failure which is not ntercepted by VMware ESXi.

Regards,
Patrick
Best regards,
Patrick
Steve_Paine
Occasional Contributor

Re: P4500 Multi-Site SAN iSCSI Initiator problems

Thanks guys, all usfull info...The hardware platforms are HP ProLiant DL380G6 servers with NC522SFP+ 10GB adapter cards. Each server has 8 other NIC's for kernal, vMotion and public access to VM's. All in seperate VLANs.

When we pull the iSCSI NICs the GUI shows them as down. on the vswitch.

The VMware NIC & Switch display commands also show the NICs down.

Its like we have a Zombie Virtual machine running....Would this happen if attached to fibre channe??