Re: Robustness
Moin Kevin Blackham,
> On another note, our continuing problems with NFS on Linux (to our NetApp or
> between machines) has been less than impressive. Processes are constantly
> blocked by waiting NFS transactions. Does anyone have suggestions for
> improving NFS performance? We've tried NFSv3... no help.
NFS is stateless, and ACID transaction secure! Both kills performance,
but improve stableness of the protocol. There had been two typical
improvements of the first in the old days.
The first one was "prestostore", which was a battery buffered ram, to
act as the transaction log, for nfs to improve write performance by
using the prestostore ramdisk log for nfs ACID transactions.
The second was a bad kernel patch (2 noops) in SunOS to avoid calling
for syncronous write of the transactions.
I think both are still the most common methods. Either provide RAM-DISK
for the transactions, or drop ACID.
Bye Michael
--
mailto:kraehe@copyleft.de UNA:+.? 'CED+2+:::Linux:1.2:13'UNZ+1'
http://www.xml-edifact.org/ CETERUM CENSEO MSDOS ESSE DELENDAM
Reply to: