[Orca-users] orca re-processing older files
Attila Mezei-Horvati
attila_mh at yahoo.com
Thu Mar 16 06:23:00 PST 2006
> Orca tries to be pretty smart about not re-reading
> files it's already read. It
> keeps track of the timestamp and size of the file,
> so if neither of these
> change, it shouldn't reread the whole thing.
>
> It will read the first line of the file to learn
> which columns it has though.
>
> Are you seeing noticable slowdown because of this?
>
I am running Orca on some of the servers since October
2004. At this point even if I add one day worth of
logs I need to wait hours to have Orca finished. I am
transferring files with rsync which as I know does not
change the timestamp or size. It just uploads the
differences which is usually the new file. I can see
in the log that orca reads through every file:
Read 288 data points from
`/.../percol-2004-12-09.bz2'.
Read 288 data points from
`/../percol-2004-12-10.bz2'.
Since it reads only 288 data points I would guess that
indeed it reads only the first lines. Maybe I should
not keep all the logs in that place? I am just worried
the graphs would change if I remove older files.
thanks,
Attila
__________________________________________________
Do You Yahoo!?
Tired of spam? Yahoo! Mail has the best spam protection around
http://mail.yahoo.com
More information about the Orca-users
mailing list