Re: rollover translator

[Harley D. Eades III]

 "Alfred M. Szmidt" <ams@kemisten.nu> writes:
 
>> Hey,
> > 
> > I got this funky idea today, but I'm not sure what would be the best
> > way to implement it.  What it would essentially do is provide a
> > "circular file", so when you get to the end of the file you start from
> > position 0.  With the nice option that if you do roll over, you can
> > clear the all stored contents and re-start with a "blank file".
> > 
> > This is very useful for if you are dumping lots of log info and you
> > are only interested in the last few billion lines.
> > 
> > Anyway, what I'm unsure of is how you should set the actual
> > translator, one way which should be trivial to do is to make it act
> > like a "link".  So you would do
> > 
> > settrans -ac /log /hurd/rollover 100M /real-log
> > 
> > And if you write more then 100M to /log, it would start from position
> > 0.  That way anything beyond 100M in /real-log would be intact.
> > 
> > Another way would be just to skip the whole /real-log story, and all
> > read/writes would happend to the underlying node.
> > 
> > Well, thats all folks...
I think this is a good idea.  This will help with managing large data holding
log files.  It seems quite efficient disk space would not grow so the hassle 
of going into your log store would be no more :). Just my two cents.

hde