Re: Performance problems with NFS connected DB server ( to NetApp filer)

Roy L Butler <[email protected]> Wed, 06 Jun 2007 10:05:06 -0700
Newsgroups gmane.comp.db.maxdb
Message-ID <[email protected]>
> What do each of these "net*" parameters represent?  TIA.
> 
>> We use the following option and decreased our backup time by 30 %:
>>
>> net.core.rmem_max=8388608
>> net.core.wmem_max=8388608
>> net.core.wmem_default=8388608
>> net.core.rmem_default=8388608
>> net.core.netdev_max_backlog=4096
>> net.ipv4.tcp_rmem = "4096 87380 4194304"
>> net.ipv4.tcp_wmem = "4096 87380 4194304"
>> net.ipv4.tcp_mem = "4096 87380 4194304"

Download the Linux kernel source and look in:

	./Documentation/networking/ip-sysctl.txt

They're basically increasing the networking stack buffer sizes and 
backlog in order to improve performance an a high-bandwidth, low-latency 
network.


Roy

-- 
MaxDB Discussion Mailing List
For list archives: http://lists.mysql.com/maxdb
To unsubscribe:    http://lists.mysql.com/[email protected]