Skip to content

Categories:

ESXi Advanced Features

To quickly display specific ESXi advanced settings, use the Get-AdvancedSetting command:

Get-AdvancedSetting -Entity (Get-VMHost) -Name [,...]

However, the ESXi host name will not be returned, so you need to pipe and select the required fields, as follows:

DataMover.HardwareAcceleratedMove,DataMover.HardwareAcceleratedInit | Select Entity,Name,Value | Sort-Object -Property Entity | Format-Table

Posted in Uncategorized.

Tagged with .


Frequent NFS APDs after upgrading ESXi 5.5 to U1

If you’re using NFS storage and have not yet upgraded to 5.5 Update 1 – DON’T DO IT YET!

VMware have just released KB2076392 stating that there is a known issue affecting ESXi 5.5 U1 hosts with NFS storage. Here’s an excerpt from the KB:

“When running ESXi 5.5 Update 1, the ESXi host frequently loses connectivity to NFS storage and APDs to NFS volumes are observed. You experience these symptoms:

  • Intermittent APDs for NFS datastores are reported, with consequent potential blue screen errors for Windows virtual machine guests and read-only filesystems in Linux virtual machines.
    Note: NFS volumes include VSA datastores.
  • For the duration of the APD condition and after, the array still responds to ping and netcat tests are also successful, and there is no evidence to indicate a physical network or a NFS storage array issue.
  • The NFS storage array logs and traces also do not indicate any evident issue, other hosts not running ESXi 5.5 U1 continue to work and can read and write to the NFS share without issue.”

Posted in Issues.

Tagged with , , .


Veeam Backup & Replication 8.0 to offer NetApp Snapshot support

Veeam have just announced that their upcoming v8 release of Backup & Replication will provide full support for NetApp integration:

  • Fast backups: back up VMware VMs from NetApp storage snapshots up to 20 times faster
  • Quick recovery: easily recover individual items from NetApp Snapshot, SnapMirror and SnapVault
  • Improved protection: improve DR protection by creating instant, secondary backups from NetApp snapshots

Veeam Backup & Replication v8 will be generally available in the second half of 2014.

For more info see http://go.veeam.com/v8-netapp

Posted in Uncategorized.


Understanding vSphere Active Memory

Here’s an interesting blog post on what ‘active’ memory is – it’s probably not what you expect:

http://blogs.vmware.com/vsphere/2013/10/understanding-vsphere-active-memory.html

The conclusion is that active memory only shows what memory pages have been touched over the last sampling peiod (20 seconds in vCenter, 5 secs in ESXTOP) – it does not describe if they are the same memory pages or different memory pages between sampling periods.

Therefore active memory cannot be used as a “rightsizing” counter on its own. In-guest memory counters are the most accurate and should be used in combination with vSphere memory counters.

Posted in Uncategorized.

Tagged with , , , .


ESXi 5.x PSOD and data corruption with E1000 vNIC

This is the top VMware support issue for November 2013, so it’s a serious bug, although we’ve not seen it yet here. The E1000 series vNICs can cause PSOD and data corruption issues in ESXi 5.x. See this article for more info:

http://kb.vmware.com/kb/2059053

Posted in Issues.

Tagged with , , .


VMware Top 20 Support Articles

This interesting VMware support page is updated every month with the top 20 support articles:

http://blogs.vmware.com/kb/top-20?

Posted in Uncategorized.

Tagged with , , , .


How SRM Issues SRDF Commands

This shows how SRM 5.1 issues commands to manage EMC SRDF disks during planned recovery and disaster recovery.

SRM Planned Recovery

SRM Event Site Command Prod Device Prod State DR DR state RDF Pair State
Normal state Prod R1 RW R2 WD Consistent
5.1.2 Change storage to read only Prod Suspend -force R1 RW R2 RW Suspended
Prod Write disable R1 R1 WD R2 RW Suspended
  1. Sync storage – disable consistency protection
Prod Disable R1 WD R2 RW Suspended
  1. 8. Change Recovery Site Storage to Writeable
DR Resume -force R1 WD R2 RW Invalid
DR Disable R1 WD R2 RW Invalid
DR Failover –force R1 WD R2 RW Failed over
Reprotect

  1. Configure Storage to Reverse Direction
DR Swap R2 WD R1 RW Suspended
DR Resume –force R2 WD R1 RW Consistent
DR Enable R2 WD R1 RW Consistent

SRM Disaster Recovery when links cut

Event Site Command Prod Device Prod State DR DR State RDF Pair State
Pre-outage R1 RW R2 WD Synchronized
*Links cut R1 RW R2 WD TransIdle (ASYNC)
Run SRM Disaster Recover

  1. Change Recovery Site Storage to Writable
DR Failover –Immediate -Force R1 RW R2 RW Partitioned (SYNC)
*Links re-established SRM cannot continue plan when in split state! R1 RW R2 RW Split
Manually power down R1 VMs and set R1s to WD Prod Power down VMsWrite disable R1 R1 WD R2 RW Failed Over
Run Disaster Recovery again DR Will complete with no changes
Reprotect

  1. Configure Storage to Reverse Direction
DR Swap R2 WD R1 RW Suspended
DR Resume –force R2 WD R1 RW Consistent
DR Enable R2 WD R1 RW Consistent

Posted in Uncategorized.

Tagged with , , .


IP Hash-Based Load Balancing and EtherChannel

On a standard vSwitch, IP Hash (the out-IP policy) uses the source and destination IP addresses to calculate a hash.

It is the only policy which allows the VM to access two physical adapters without having two virtual NICs.

Requirements are:

  • all physical network adapters must be connected to the same switch
  • OR if connecting to two switches, link aggregation across physical switches must be used to ‘stack’ the switches
  • the switch must be configured for link aggregation (802.3ad) / EtherChannel
  • for EtherChannel to work across two physical Cisco switches, MEC (Multichassis EtherChannel) must be used on 6500 or Virtual Port Channel (vPC) on Cisco Nexus 7000. See here for more info.
  • all physical NICs must be set to active, instead of having some active, some passive

Note: although it can more evenly spread the transfer traffic for a single VM, it does not provide a benefit for large data transfers occurring between the same source and dest IPs.

See also:

Host requirements for link aggregation for ESX and ESXi (1001938)

Sample configuration of EtherChannel / Link Aggregation Control Protocol (LACP) with ESXi/ESX and Cisco/HP switches (1004048)

Troubleshooting IP-Hash outbound NIC selection (1007371)

Posted in Uncategorized.

Tagged with , , , .


Veeam ONE Database performance issues

I was seeing issues where Veeam ?Reporter was not publishing scheduled reports and performance was poor.

This was a very simple fix!  It was because the SQL transaction log database was set to start at 1MB and auto-grow by 10%. This was not big enough, so we set initial size to 50MB and auto-grow by 10MB. :-)

Posted in Issues.

Tagged with , , , .


Veeam ONE Data Collection VMware Job Timeout

Veeam ONE Reporter job logs the following errors (Configuration -> Sessions):

There were errors while collecting VM files: There were errors in VM files collection process: Failed to retrieve VM files due to collection timeout.

This is due to timeouts within Veeam ONE as per http://www.veeam.com/kb1653

In our environment:

  • Changed reg key HKLM\SOFTWARE\Veeam\Reporter Enterprise\GetVmSnapshotTimeOut to 7200 decimal = 120 mins
  • Added HKLM\SOFTWARE\Veeam\Reporter Enterprise\VIFilesDatastoresCountByTime to DWORD 1
  • Added HKLM\SOFTWARE\Veeam\Veeam ONE Monitor\Service\HistQueryTimeoutSec REG_DWORD to (36000)

Posted in Issues.

Tagged with , , .