Preston Boyington wrote: > currently my two biggest bottlenecks are with the networking and the > MTA. Could you explain further how the MTA on your laptop is the bottleneck. That seems very strange to me and I can't think of how it would be a bottleneck. > i installed "ifplugd" and set a faster timeout for the network and > that has helped. I like ifplugd and also use it on my machine. But not for any performance reasons. It simply provides simple and nice functionality for the task that it does. > i don't really "need" a MTA for my laptop and will be > looking for some alternative. have thought about just removing Exim and > dealing with whatever dependency issues arise. i can always just look > at the log files for info if i need to. If you want to remove the MTA all together then instead of doing that you might consider installing 'nullmailer'. It "provides" mail-transport-agent and so all of the dependencies will be satisfied. I think that would be better than forcing through the dependency problems by removing the mta entirely. Also, if you believe that your mta is causing you problems and are to the point of removing it then I suggest that you simply stop it first and see if that is really the issue. /etc/init.d/exim stop # old version 3 /etc/init.d/exim4 stop # new version 4 After it is stopped then it is almost as if it were not installed. That would be a good simulation to test the theory that it is a bottleneck. If whatever problems were observed remain then it probably isn't the mta. But if that test fixes the problem then it would be a good indicator that the mta is the issue. (I doubt that it is though.) > >I altered the link from /bin/sh to point to /bin/dash instead of > >/bin/bash but I'm not sure if I got a big improvement. It seems > >reasonably fast though. :-) > > i have been using zsh recently and it has worked pretty well so far. "worked pretty well" is great but unfortunately not something that can be measured or quantified. When talking performance it is numbers and data points that really speak clearly. If possible anyone who is trying to do system speedups should measure a data point both before and after. Otherwise random noise into the system can really cause confusion and misinterpretation of the results. I caution this because in past experiences I have seen people make optimization changes without data that actually had the opposite effect and slowed things down! > compiling my own kernel. the generic one is great, but with one > specific to the machine i should pick up some more speed. What cpu do you have in your laptop? If i686 install the 686 optimized package. If amd64 then install that one. And so forth. Those are optimized for their architecture. linux-image-2.6-686 linux-image-2.6-amd64 The ones who lose out are folks like me running systems with older cpus such as the K6 which no longer have optimized kernels available specificially for their architecture in Debian. I have dropped back to using the compatible 486 kernel on that system. It is not tuned but is good enough for my purposes. It is not my main desktop anymore and now functions in a supporting server role. Bob
Attachment:
signature.asc
Description: Digital signature