MaxDisplays configuration option

[AG]

From decades in the field, I'd say It's pretty weird. I've not
personally seen anything approaching that number of clients on a
single server in..... well, not since I worked with Multics back in
the 1980's. Dozens on a robust system, yes. One thousand?

We do have some boxes with concurrent ssh-users in the lower 3-digit
range. But in general this seems to be rare, especially since software
is often unprepared for and untested in that amount of activity (see
e.g. https://github.com/systemd/systemd/issues/1961).
For the number of X displays, there was never any issue, usually since
CPU and memory resources run out long before you run out of display
numbers. Users just pick another box or their laptop, if applications
are slow, so the number of X displays is self-limiting ;)
Ciao,
Alexander Wuerstlein.

I should clarify, these aren't systems that have users shell access to.
It's an X11 forward only, used as a chokepoint into a segregated network.

Anyway, I will clean up the patch and add documentation and then report
back. I've already created an entry on the bugtracker in case anyone wants
to follow along.

https://bugzilla.mindrot.org/show_bug.cgi?id=2580

Thx,
A

[AG]

FYI- I have attached a finalized version of the X11MaxDisplays patch to the
mindrot bugzilla.

https://bugzilla.mindrot.org/show_bug.cgi?id=2580

The issues with variable naming, whitespace and other style issues are
addressed and
I've added documentation to sshd_config

I appreciate any feedback, but I think this is in pretty good shape to be
considered.

Thx.

A

Also, I'm curious if this issue has ever come up before? Is it really

that

strange of a case?

From decades in the field, I'd say It's pretty weird. I've not
personally seen anything approaching that number of clients on a
single server in..... well, not since I worked with Multics back in
the 1980's. Dozens on a robust system, yes. One thousand?

We do have some boxes with concurrent ssh-users in the lower 3-digit
range. But in general this seems to be rare, especially since software
is often unprepared for and untested in that amount of activity (see
e.g. https://github.com/systemd/systemd/issues/1961).
For the number of X displays, there was never any issue, usually since
CPU and memory resources run out long before you run out of display
numbers. Users just pick another box or their laptop, if applications
are slow, so the number of X displays is self-limiting ;)
Ciao,
Alexander Wuerstlein.

[Nico Kadel-Garcia]

Post by AG

From decades in the field, I'd say It's pretty weird. I've not
personally seen anything approaching that number of clients on a
single server in..... well, not since I worked with Multics back in
the 1980's. Dozens on a robust system, yes. One thousand?

We do have some boxes with concurrent ssh-users in the lower 3-digit
range. But in general this seems to be rare, especially since software
is often unprepared for and untested in that amount of activity (see
e.g. https://github.com/systemd/systemd/issues/1961).
For the number of X displays, there was never any issue, usually since
CPU and memory resources run out long before you run out of display
numbers. Users just pick another box or their laptop, if applications
are slow, so the number of X displays is self-limiting ;)
Ciao,
Alexander Wuerstlein.

I should clarify, these aren't systems that have users shell access to.
It's an X11 forward only, used as a chokepoint into a segregated network.
Anyway, I will clean up the patch and add documentation and then report
back. I've already created an entry on the bugtracker in case anyone wants
to follow along.
https://bugzilla.mindrot.org/show_bug.cgi?id=2580
Thx,
A

*Oh*. OK, that suddenly makes a great deal more sense.