I agree with TobiSGD's answer.
Some extra detail:
Quote:
Originally Posted by SilvioGesell
But this morning I have seen 58 KB swap used. Swapiness is 60. So either the RAM has grow up this night or .. I dont know.
|
Apparently cache use grew. It tends to gradually do that as a system with excess ram (compared to workload) stays up.
If you saw a big change over night, that tends to mean some scheduled task does a lot of file I/O, which is probably normal.
Any anonymous memory of processes that had been idle for a long while before that burst of file I/O will be swapped out. When you resume using those processes again, there will be a moment of poor responsiveness as those pages come back in. So far as I know, there isn't a good way to tune that sub optimal behavior out (but people try a lot of bad ways, doing more harm than good). It generally is NOT a big enough problem to care about.
Generally some anonymous pages of some services are stale data (typically left over from the service startup). Once cache usage has been high enough (whether gradually or due to a big burst of file I/O) those pages stay in swap, because nothing ever needs them in ram.
Some of the 58MB you see is stale data with no reason to come back into ram. Maybe some of it belongs to some service that hadn't been touched again after whatever file I/O kicked. 58MB is small enough that the delay to bring it back will be trivial if whatever service owns that data resumes work.