After a few days, I notice that my machine slows down appreciably and there is often a great deal of thrashing. When I check the statistics using “top”, I find that a certain amount of swap space has been used. When I shut down all of the user applications, the amount of swap space doesn’t go back to zero and I find that I don’t really get rid of thrashing until I reboot. What is happening here? Also, does hard disk access hog the processor all to itself? I would have thought that all processes had to give up the processor and do a context switch when their slice was up. I am using Feisty Fawn, btw.
Never had that problem with Fiesty or the current Hardy, or any of them, come to that.
You would need to post your hardware for your question to be meaningful but even better, direct it to the Ubuntu Forums where they have more intimate knowledge and far more Ubuntuists waiting to take your call.
They are as friendly and helpful as Dopers but will need your hardware specs to be able to advise you.
If it were Hardy, I’d say you may have caught it indexing but it isn’t anything that happens with the same regularity or intensity as Vista appears to do.
Yeah, what’s up with that? I just caught my laptop doing this today (scrollkeeper-update) for the first time and it pissed me off that there didn’t seem to be a way to disable it. I renamed the binary and made a symbolic link to /bin/true, which should take care of that…grrrr.
When you say “thrashing”, are you using the term in the technical sense? If so, I suspect you have a severely under-memoried computer for whatever application it is you’re attempting to run. But I suspect not, because of the following:
See, that’s the definition of thrashing; the processor is spending all it’s time context switching rather than doing useful computation.
It could very well be that updatedb or scrollkeeper is running. Yes, they’ll heavily tax (i.e., “thrash”, in the non-technical sense) the CPU. But note that this is not simply “disk access”, which nowadays is most likely performed using DMA (direct memory access). Rather, there’s a bunch of processing going on (either indexing, in the case of scrollkeeper, or…ummm…updating the system database in the case of updatedb). I’m pretty sure on Ubuntu there’s an /etc/updatedb.conf, in which you might choose to change the NICE setting.
Also, your swap should never go down to zero (assuming that you’re referring to cached, not active memory).
Hardy did a bit of indexing when the machine was idle, nothing spectacular, but hasn’t done anything noticeable since the first day.
It is left on all the time and is often (mostly) unattended.
You may want to check out this article, it gives a basic description of what can be slowing down your system as well as a helpful guide to the NICE commands and their arguments.