TR-3428 Best Practice recommendation: Esxcfg-advcfg –s 120 /Net/TcpIpHeapMax

[finn]

Esxcfg-advcfg –s 120 /Net/TcpIpHeapMax

What is the justification for this setting as this changes the RAM allocation within the VMKernel to TCP.

Thank you!

[amiller_1]

Needed I believe if you do go beyond 8 NFS datastores....recommended by VMware actually when adjusting the Max NFS Datastores from 8 to 32.

http://kb.vmware.com/selfservice/microsites/search.do?language=en_US&cmd=displayKC&externalId=2239

Edit: also helps with removing cosmetic error messages (which is an excellent idea....never good to just ignore error messages when you can resolve them (or if cosmetic remove them).

http://communities.vmware.com/thread/170499

[RichardSopp]

The documented heap size reflects the proposed 32 NFS datastore configuration in the document. Setting the maximum number of datastores when configuring the system involves an ESX host reboot further down the line to make the change.

What I think is missing from the current document is what are the heap size and other settings if you don't want to increase all the way to 32 NFS datastores but want something larger than 8.

[amiller_1]

Hmm...to be honest, I'd just say that if 8 isn't sufficient it makes sense to jump to 32 period.