At Wit's End on system freezes

• Messages sorted by:[ date ][ thread ][ subject ][ author ]

Gert Doering <gert@greenie.muc.de> wrote:
> I *wrote* mgetty to make bidirectional port usage work robust and

I can tell that, using mgetty for > 3 years on many different
Linux configuration, and having using it also on very
catastrophic hardware (BULL DPX SYSVR2 and MIPS M/120 SVR4)
that it *just works*.

I had *no single failure* in 3 years on Linux which could be
attributed to mgetty. I had three failures, one due to
the modem overheating, and two since I upgraded the ZyXEL 2864I
to multi-MSN firmware in may 98. The last two failures have
locked the ISDN TA :)

Of course, I have seen unreliable mgetty setups. Typically when
using symlinks, incorrect paths for LOCK files, different
speeds for dialin-dialout bad cables, etc.

But usually, those get diagnoticized very fast. I also like the
mgetty feature that when the modem answers, the system *must*
be in good shape (ie, mgetty DOES the ATA).

One things that REALLY makes modem behave better is the
hard-reset on DTR drop option that some have.

Another, with Linux, is to *only use the ttyS* interface for
in *and* out. Anyway, the cu* interface is going to disappear
quite soon now, and it doesn't interact correctly with
mgetty's way.

>  - compiling linux 2.0.x kernels with C compilers different to gcc 2.7.2.3
>    (egcs for sure breaks 2.0 kernels [because kernel ASM and egcs ASM just
>    don't work together, I'm not blaiming anyone, just stating facts]).

2.0 kernels should NEVER be compiled with egcs. This is quite clear
by reading the linux-kernel mailing-list.

> Mgetty cares, but all it will do if the modem is hosed is to complain 
> into the log files (possibly filling up your disk), but never "freeze 

yes, this happens, especially with overheating the ZyXEL modem :)
Past 35 degrees is quite bad for hardware (the server also filled
the logs with ``keyboard errors'' --- funny, isn't it ?).