|
|
Jon Bendtsen wrote:
on my client computer, enabling comp-lzo used alot of CPU. That's unusual -- LZO typically can compress data at about 1/12 the speed of a raw memcpy() and decompress at 1/3 memcpy() speed. Unless you're transferring *massive* amounts of data, it shouldn't make a measurable impact (except inasmuch as the larger amount of throughput via the tun device causes increased overhead elsewhere). I know how to start tcpdump, and maybe ethereal. But I don't know what I'm looking for. Can you give me a hint? See if you have patterns where large packets sent on one side show up as multiple smaller packets on the other, are rejected en route such that they need to be resent as fragments by the endpoint itself, or simply disappear. OpenVPN's mtu-test will try to automatically determine what the largest packet size that actually goes through is; you can then use its values to update your config file. That said, when you have a specific issue like this, "anything unusual" is a good thing to look for. Is the database client doing continuous communication with the server during the startup phase, or is there a wait state somewhere where it pauses? If you find a spot where it's waiting for something to happen (which isn't very unusual at all), try to determine exactly what that is.
Warning: require_once(../../../archive_common.php) [function.require-once]: failed to open stream: No such file or directory in /home/openvpn/domains/openvpn.net/public_html/archive/openvpn-users/2006-04/msg00085.html on line 201 Fatal error: require_once() [function.require]: Failed opening required '../../../archive_common.php' (include_path='/usr/local/lib/php') in /home/openvpn/domains/openvpn.net/public_html/archive/openvpn-users/2006-04/msg00085.html on line 201 |