Re: [Mailman-Developers] Security
On Mon, 23 Oct 2000 20:09:01 -0700 (PDT) Dan Mick <Dan.Mick@West.Sun.COM> wrote:
If I post to a specific list, or approve a held post for that list, there is an 80% chance that this will crash the machine (compleat lock, no interrupts, no useful log entries).
What do you mean "crash the system"? Panic and coredump? Hang? Crash mailman/python and everything else is still up?
As stated: compleat lock, no interrupts being serviced, no log entries. Basically: utterly indistinguishable from a powered off machine other than by power consumption. No KB, no met, no stack, no video (well, a slight grayish diffuse glow actually), nuttin'.
If panic/coredump, some time spent in cursory kernel coredump analysis is warranted to point a finger at a culprit. Being a kernel developer, I'm biased, but when people say "crash the machine" that's what I think of. Is that what you mean?
Nope. No panics, no cores, just one big wedgie.
These are among the reasons this whole thing fits the description of "odd".
-- J C Lawrence Home: claw@kanga.nu ---------(*) Other: coder@kanga.nu http://www.kanga.nu/~claw/ Keys etc: finger claw@kanga.nu --=| A man is as sane as he is dangerous to his environment |=--
"JCL" == J C Lawrence <claw@kanga.nu> writes:
JCL> As stated: compleat lock, no interrupts being serviced, no
JCL> log entries. Basically: utterly indistinguishable from a
JCL> powered off machine other than by power consumption. No KB,
JCL> no met, no stack, no video (well, a slight grayish diffuse
JCL> glow actually), nuttin'.Here's another thought. Are you running on an SMP machine? I believe there have been reports of older versions of Python having problems when threads and SMPs are mixed. Have you enabled the threading delivery in SMTPDirect? Turning that off, or upgrading to Python 2.0 (which I believe has fixed the problems I vaguely recall) might do the trick.
-Barry
On Mon, 23 Oct 2000 23:34:18 -0400 (EDT) barry <barry@wooz.org> wrote:
"JCL" == J C Lawrence <claw@kanga.nu> writes:
JCL> As stated: compleat lock, no interrupts being serviced, no log JCL> entries. Basically: utterly indistinguishable from a powered JCL> off machine other than by power consumption. No KB, no met, no JCL> stack, no video (well, a slight grayish diffuse glow actually), JCL> nuttin'.
Here's another thought. Are you running on an SMP machine?
Yes.
I believe there have been reports of older versions of Python having problems when threads and SMPs are mixed. Have you enabled the threading delivery in SMTPDirect? Turning that off, or upgrading to Python 2.0 (which I believe has fixed the problems I vaguely recall) might do the trick.
What I find most odd is taht I've processed several hundreds of thousands of messages thu this exact setup without fail (200 day uptime, half dozen lists, etc) and yet now its unhappy roughly 80% of the time. Something lese is odd I suspect.
-- J C Lawrence Home: claw@kanga.nu ---------(*) Other: coder@kanga.nu http://www.kanga.nu/~claw/ Keys etc: finger claw@kanga.nu --=| A man is as sane as he is dangerous to his environment |=--
participants (2)
-
barry＠wooz.org -
J C Lawrence