This post is about bugs you could be concerned about. I will try to keep it up-to-date.
ESXi 6.0 U2 runs in PSOD when VMs running on Version 11, uses VMXNET3 NICs and have Large Receive Offload (LRO) enabled. On PSOD you can see these errors:
Vmxnet3VMKDevRx. To resolve this issue, either disable hardware LRO on ESXi hosts (advanced setting:
0) or revert VMs to previous version. See KB for more information.
Again a bug in CBT that causes inconsistencies in incremental backups. Happens in ESXi 6.0
To solve the problem, install newly released fix. More information you can find here. Veeam comments that resetting CBT is mandatory!
When running a host on ESXi 5.5 U3 you can have huge problems with snapshot consolidataion tasks:
- VM failure after consolidation
- VM outage after consolidation.
vmware.log you can find the singnal error:
Caught signal 11.
To resolve this issues you can either to to U3a or revert to an older Update-Version. For more information click here.
When trying to migrate storage from a VM (powered-on or powered-off) with CBT enabled from a ESXi 5.5 to a ESXi 6.0 host this may fails! Keep in mind when planing your upgrade from 5.5 to 6.0.
More information see here.
After upgrading to ESXi 6.0 U1 you will have problems to create or recompose pools using linked clones with View 6.1 and older versions. You see an error like: View Composer DiskFault: Disk customization failed due to an internal error in View Administrator.
If you are unable to login to vSphere Web Client, it could just be that the service is running more than one year without restarting it. You can suffer from this issue if you
- Cannot login to Web Client
- Cannot login to vCenter using vSphere Client
- Get an error like:
Cannot connect to vCenter Single Sign On server. https://(SSO_Server):7444/ims/STSService.
during login to Web Client.
The solution to this is to restart vSphere Web Client Service. More infos here.
When creating a quiesced snapshot in ESXi 4.x or 5.x for Windows Server starting with 2003 you may get duplicated disks. You can also get these events in Windows event viewer:
- ID: 50, NTFS, Warning
- ID: 57, NTFS, Warning
- ID: 137, NTFS, Warning
- ID: 140, NTFS, Warning
- ID: 157, Disk, Warning.
To solve this problem, install proper patches you can find here.
If you upgraded Windows OS on a machine were vCenter components are installed, services can have problems to start during system startup. This is because dependencies are not working any more because Microsoft renamed services in Windows Server 2012 R2. To correct dependencies use this article.
If you use Broadcom BCM5719 or BCM5720 NICs or any other NIC that uses the tg3 driver you can have this issue. The NIC stops working. So you can have network outages or just loss of redundancy. More information can be found here.
To verify whether jumbo frames works from host to host or storage on all devices you can use
vmkping (using parameter
-d). Unfortunately there is a problem in 5.5. You can get an error like
sendto() failed (Message too long). This can happen when you change the MTU size of an vmkernel port multiple times. This can be solved by re-creating the port and do not change the size multipe times. More information can be found here.
During very high network traffic bursts you can see package loss with vmxnet3 vNICs. This can occur on 4.x and 5.x hosts, caused by shortage of buffer space. More information here.
A huge problem, that could hit many. On 5.5 U1 hosts all path down (APD) condition can occur periodically on NFS volumes. In the meantime a fix is available! More information here.
Heartbleed. Here are more information on affected products and fixes
Problem with vNIC e1000 and e1000e could cause a PSoD on 5.x hosts. Patches and workarounds are available. More information here.
Did you know this one: some time ago, a bug convinced 3.5 hosts to be without license anymore. This happens exactly on August 12, 2008. Read more about this one here.