Bug #48853 Details of SendBufferMemory configuration contains errors
Submitted: 18 Nov 2009 2:22 Modified: 16 Dec 2009 11:39
Reporter: Jonathon Coombes Email Updates:
Status: Closed Impact on me:
None 
Category:MySQL Server: Documentation Severity:S3 (Non-critical)
Version:5.1-ndb-7.0+ OS:Any
Assigned to: Jon Stephens CPU Architecture:Any
Tags: cluster, documentation, manual, ndb, sendbuffermemory

[18 Nov 2009 2:22] Jonathon Coombes
Description:
The documentation regarding the send buffers and associated options has errors in it. The documentation is at:

http://dev.mysql.com/doc/refman/5.1/en/mysql-cluster-config-send-buffers.html

The ReservedSendMemoryBuffer option does not appear to be used at all in the code?

The OverLoadLimit has the default listed as 0, which gives indication of "unlimited". However, the code shows that the default is calculated as:

conf->tcp.sendBufferSize*4/5

The whole paragraph on the SendBufferMemory is quite confusing. Breaking it down by sentence, we get:

SendBufferMemory denotes a hard limit for how much memory, out of the  TotalSendBufferMemory pool, may be used by a single transporter. This seems to make more sense.

"However, the sum of TotalSendBufferMemory for all configured transporters may be greater than SendBufferMemory."  This is incorrect. Based on the principle that TotalSendBufferMemory is the maximum amount per node, you cannot get a sum of them. It makes more sense to be reversed to read "However, the sum of SendBufferMemory for all configured transporters on the node, may be greater than the TotalSendBufferMemory for that node".

How to repeat:
N/A

Suggested fix:
N/A
[16 Dec 2009 11:39] Jon Stephens
Thank you for your bug report. This issue has been addressed in the documentation. The updated documentation will appear on our website shortly, and will be included in the next release of the relevant products.