VMware Cloud Community
Abadon
Enthusiast
Enthusiast
Jump to solution

Storage migration fails with the error: Could not complete network copy for file path to file

Hello,

I have two ESXi 5.5 hosts with local storage and VMware vCenter appliance which manage them. Everything worked fine till this moment.

Now when I try to make host and storage migration of working or power off virtual machine I receive the error message:

Could not complete network copy for file /vmfs/volumes/..../path/to/file.vmdk

The situation is the same when I try to deploy VM on ESXi-2 from template located on ESXi-1.

Everything with network setup is correct. I have ping between ESXi hosts and between vCenter and ESXi hosts. No firewalls on ESXi hosts or vcenter. No physical firewalls between hosts.

The network latency is under 1ms. No packet losses.

I read this article VMware KB: Storage migration fails with the error: Could not complete network copy for file and tried solutions explained there, but results are the same.

Do you have ideas where the problem can be?

0 Kudos
1 Solution

Accepted Solutions
Abadon
Enthusiast
Enthusiast
Jump to solution

Hello guys,

I finally found the problem.

The ESXi servers are configured with MTU 9000, but the switch where hosts are connected had a documented bug with Jumbo frames. I found the problem after find this article http://kb.vmware.com/selfservice/microsites/search.do?language=en_US&cmd=displayKC&externalId=100373...

I execute command from ESXi-1:

vmkping -d -s 8972 192.168.10.12  where 192.168.10.12 is ip address of ESXi-2

The result was:

3 packets transmitted, 0 packets received, 100% packet loss

like workaround I reduce the MTU to 1500, then I patched the problematic switch.

View solution in original post

0 Kudos
8 Replies
admin
Immortal
Immortal
Jump to solution

Check if vmfs is running out of heap.. Grep for this error message in the vmkernel log

"WARNING: Heap: 2638: Heap vmfs3 already at its maximum size. Cannot expand"

0 Kudos
akblpatel
Enthusiast
Enthusiast
Jump to solution

0 Kudos
sakthivelramali
Enthusiast
Enthusiast
Jump to solution

Hi

Please ensure that the destination datastore is formatted with a block size that is large enough to support the VMDK file of the source machine.

Thanks Sakthivel R
0 Kudos
Abadon
Enthusiast
Enthusiast
Jump to solution

In /var/log/vmkernel.log I cannot find such warning/errors.

I run:

~ # grep "Heap" /var/log/vmkernel.log

~ # grep "heap" /var/log/vmkernel.log

but do not receive any result.

0 Kudos
Abadon
Enthusiast
Enthusiast
Jump to solution

The both data stores are absolutely the same, because the hardware is 100% identical.

File-system type: VMFS 5.60

Block size: 1MB

I made copy via SSH of VM template from Host-1 to Host-2.

Now I can deploy new VMs on Host-1 from Template located on Host-1 or VMs on Host-2 from Template located on Host-2.

That's why I think that the problem is not into storage.

0 Kudos
Mahiee
Enthusiast
Enthusiast
Jump to solution

Hi - Just take a insight into it if helps: http://kb.vmware.com/selfservice/microsites/search.do?language=en_US&cmd=displayKC&externalId=203849...

Apart can you confirm if those two hosts are part of any cluster or standalone hosts.

And also have you ever moved any host or Data files moved as you are trying now.

0 Kudos
Abadon
Enthusiast
Enthusiast
Jump to solution

Hi,

I started with link for provided article. In my first post I put the same link, but due unknown for me reason the forum wrote only:

I read this article VMware KB: Storage migration fails with the error: Could not complete network copy for file and tried solutions explained there, but results are the same.

Anyway. I've done these steps on both hosts but the problem persists.

The both hosts are part of one Cluster.

Do you think that if I remove the hosts from the cluster, the problem will be resolved?

Best regards!

0 Kudos
Abadon
Enthusiast
Enthusiast
Jump to solution

Hello guys,

I finally found the problem.

The ESXi servers are configured with MTU 9000, but the switch where hosts are connected had a documented bug with Jumbo frames. I found the problem after find this article http://kb.vmware.com/selfservice/microsites/search.do?language=en_US&cmd=displayKC&externalId=100373...

I execute command from ESXi-1:

vmkping -d -s 8972 192.168.10.12  where 192.168.10.12 is ip address of ESXi-2

The result was:

3 packets transmitted, 0 packets received, 100% packet loss

like workaround I reduce the MTU to 1500, then I patched the problematic switch.

0 Kudos