On 02/24/10 02:04 PM, Shawn Walker wrote:
On 02/24/10 03:59 PM, Haik Aftandilian wrote:
On 02/24/10 12:38 PM, Shawn Walker wrote:
On 02/24/10 02:30 PM, Haik Aftandilian wrote:
I'm running build 132 from the /dev repo on a dual core P8400 2.26GHz
with 2GB of RAM and my system became slow and then almost totally
unresponsive while the pm-checkforupdates script was running. It seemed
like all applications' memory had been paged to disk because I could
hear the disk running and the desktop environment was very
slow--unresponsive to keyboard commands and very slow to switch
windows.
prstat showed the process CPU utilization was over 30%. I used mdb to
save a core file which turned out to be 622M in size. I've attached the
stack trace. Does this sound like a known bug? I didn't see a release
note about it. Would more information be needed beyond the stack trace?
If you're on build132, then the update check for newer builds (133+) is
very expensive. This is a known issue. Once you upgrade to build133,
you'll see the memory usage drop significantly for future update checks.
OK, good to know. Thanks. In that case I won't file a bug. I wasn't sure
if you meant that there is already a bug filed for this on
pm-checkforupdates or if this is just a one-off problem with build 132
packaging.
It's a general memory issue that a bug has already been filed for. The
reason for the memory usage is that an upgrade to 133+ from any prior
build means that the package system has to process double the number of
packges due to the giant package renaming that happened in 133.
See http://defect.opensolaris.org/bz/show_bug.cgi?id=9388
That answers my question. Thanks for the quick response.
Haik
_______________________________________________
pkg-discuss mailing list
[email protected]
http://mail.opensolaris.org/mailman/listinfo/pkg-discuss
