> > 149 nscan = (last_total_pages * > > es * (tune.t_fsflushr))/v.v_autoup; > > > > nscan can be reduced by removing memory from the > > system; by lowering > > tune.t_fsflushr (but apparently that is already > > reduced to 1 on S10 GA); or > > by increasing the v.v_autoup tunable. > > tune.t_fsflushr used to be 5 and autoup 30 for > solaris 9 by default. why was tune.t_fsflushr > reduced to 1? now fsflush is invoked more often and > it always has some pages for writing to disk because > untar has large amount of small files and the > operation takes longer than 30 seconds (skips > actually start after 30 secs of unpack, so this looks > like an exp.). > > what kind of failures require fsflush to be called as > often as every second by default?
so, I doubled up the values for both of these parameters and the symptoms seemed to disappear while unpacking but within seconds the same phenonmenon repeats. This message posted from opensolaris.org _______________________________________________ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org