When I run top, I see:
Mem: 27072K used, 100256K free, 0K shrd, 0K buff, 10584K cached Load average: 0.33, 0.25, 0.21 (State: S=sleeping R=running, W=waiting)
What does "10584K cached" mean? My system crashes once I get above
12000K cached. Is there a configuration variable I can change somewhere to stop it from crashing when cached reaches 12000K?I am running Linux 2.6.10 on powerpc.
That might not be a stable kernel. In debian we used 2.6.8 and then pretty quickly went to 2.6.18 and now in the 20s. There was a major transition between 2.6.15 and 2.6.18.
You should try a recent kernel.
What model machine do you have and what processor ? Are you sure all your RAM is really good ? You could try moving/removing a card. Small RAM modules are sometimes not tolerated well anymore (small means 16-32MB) -- if you have any of these best to try taking them out if you are having crashes.
That's about all I can say unless you tell me the particular machine and date/version/name of your distribution.
Sorry, to answer your question: cached/buffers means memory that is used to save on disk-i/o I believe. To be able to reuse libraries rather than reread them is much faster. Also not to write to disk right away. Perhaps too some video buffer if you are low on video memory.
Usually the devices can be told to read/write from/to a memory space and the cpu can go along its business if it can.
It's a custom board with 128 MB of RAM and an MPC8248 processor. Others have said to try a recent kernel. Do you have a specific reason why a newer kernel will fix the problem?
It is always best to use the latest version, because of various bugs being fixed. Anyway, he said you should try - not that it is going to fix your problem. It might be something else
g,
I don't think 2.6.10 was a stable kernel, unless the manufacturer supplied it but even then it does not look good to me. Its probably just so you can test it and know it more or less works.
depending on whether you have upgraded any other software on it then things could really be broken.
If you install a stable release and you find stuff like this happening then if I were you I would contact the manufacturer. May be there is a hardware problem, but as I said before check that your RAM is not defective !
Maybe there are kernel parameters to fix the cache size, but since that was experimental in 2.6.10, you really should get a newer kernel. Then go to some kernel gurus and ask them about setting it dynamically, if you must. But it is not so good an idea, it will slow it down badly.
Also the size of the shared memory (ipcs -m) is included into the cached number.
Regards,
Sani
Sorry, to answer your question: cached/buffers means memory that is used to save on disk-i/o I believe. To be able to reuse libraries rather than reread them is much faster. Also not to write to disk right away. Perhaps too some video buffer if you are low on video memory.
Usually the devices can be told to read/write from/to a memory space and the cpu can go along its business if it can.
ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.