[Bug 280846] Low memory freezes / OOM: a thread waited too long to allocate a page

https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=280846

Mark Millard <marklmi26-fbsd@yahoo.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |marklmi26-fbsd@yahoo.com

--- Comment #1 from Mark Millard <marklmi26-fbsd@yahoo.com> ---
Can you specify more context?

ZFS vs. UFS vs. a mix?
Type(s) of storage media in use?
Any use of tmpfs or the like?
SWAP space configuration: how much SWAP?

Output analogous to:
# dmesg -a | grep " memory "
real memory  = 2066735104 (1970 MB)
avail memory = 1990144000 (1897 MB)

And to:
# swapinfo -m
Device          1M-blocks     Used    Avail Capacity
/dev/gpt/PkgBaseSwp3p5      3584       18     3565     1%

Video hardware (and how its is configured)?

Can you tell that it looks to be leading up to
a "a thread waited too long to allocate a page"
notice before it actually happens and before
things actually hang up? Vs. is the change
sudden, going from normal to hung up to later
reporting "a thread waited too long to allocate
a page"?

Can you leave "top -Sazores" running someplace where
you can then look at or monitor top's output if you
start to notice the symptoms? What do the 2 lines
like:

Mem: 2344K Active, 1302M Inact, 404K Laundry, 320M Wired, 194M Buf, 275M Free
Swap: 3584M Total, 18M Used, 3565M Free

show? What about the process list with the larger
RES(ident) RAM use figures (top area of the process
list, via the "ores" part of the arguments to the
top command)?

Similarly for the "gstat -spod" output is generally
like?

How few seconds, minutes, hours, days, weeks, or
months can a firefox session run before the problem
occurs? (I do not know the  general scale involved
from the descriptions so far.) Is there any common
context involved across the cases of a shorter time
to the problem --that is not involved in the longer
times to having the problem?

Note that OOM kills do not necessarily kill what you
would like them to kill. They can kill the process(es)
that are allowing you to control the machine in a
normal way.

So how few "just HomeDepot" tabs can  you create and
use and get the problem in your experience so far? How
many can you create and use and not get the problem?

Are you using X11? Something else? What window manager?
And so on?

-- 
You are receiving this mail because:
You are the assignee for the bug.