trinkle@cs.purdue.edu (06/28/91)
Does anyone know why stat()ing the block /dev entry for swap
devices takes so long under SunOS 4.1.1? I wrote a trivial program
that just does a stat() on each argv argument. When run (via trace)
with /dev/sd* as arguments, it hangs for a second or more each time it
comes to a swap device (e.g. sd0b). From the activity light on the
front of the drive, it appears the kernel is reading a lot of data
from the disk while stat() is waiting. The I/O blocksize for the
device is the same as for other block devices. This does not happen
for the corresponding raw device (e.g. rsd0b), and it does not happen
for any of the other block devices. It does not happen at all under
SunOS 4.0.3.
One reason I am asking is that this behavior makes the ofiles
program very slow. Also, it seems to be a very bizarre behavior. A
workaround is trivial, just remove the appropriate /dev/sd* entry, but
I would like to know why it is happening. Does anyone have any
information that might be useful?
Daniel Trinkle trinkle@cs.purdue.edu
Dept. of Computer Sciences {backbone}!purdue!trinkle
Purdue University 317-494-7844
West Lafayette, IN 47907