[ale] memory exhausted ?

[Mark Shewmaker]

On Tue, Dec 15, 1998 at 02:33:03PM -0500, Mir Shafiqul Islam wrote:
> I get a "Memory Exhausted" when I run who. Any idea why I might getting it
> ? Here is the output of some vitals. There seems to be enough physical and
> swap available. It is RH 4.2 upgraded to 5.0 machine. Thanks.

[...]

> oispX: /users/admin/mir >> who
> who: Memory exhausted

Get the latest 5.1 sh-utils.  There's a description of the problem at
http://www.redhat.com/support/docs/rhl/rh51-errata-general.html

|Package: sh-utils 
|
|Updated: 20-Jul-1998 
|
|Problem:
|
|      (20-Jul-1998)
|
|      This update should fix problems that some users were experiencing
| with who core dumping on them after a certain number of users had
| logged in. This problem only affected some machines, and was very hard
| to duplicate, thus the time to release this fix. 
|
|Solution:
|
|      Intel: Upgrade to sh-utils-1.16-14.i386.rpm
|      Alpha: Upgrade to sh-utils-1.16-14.alpha.rpm
|      SPARC: Upgrade to sh-utils-1.16-14.sparc.rpm

Unfortunately it's listed on the 51 errata page, but not on the 5.0
errata page.  I don't know if it will work on Redhat 5.0 or not.
(I would guess you'll have the required dependencies, but I don't
really know.)

Minor rant:

I had the same problem for the longest time.  It was mentioned on
one of the linux newsgroups about a year ago, and someone had submitted
a fix to redhat in January 1998, if I remember correctly.  I had been
waiting for redhat to release their updated rpm.  (I figured it was a
small enough problem to be susceptible to bureaucratic delays of sending
code to FSF and waiting for their newer, official sh-utils or some such,
even though that couldn't have explained such a huge delay--I've not heard
of the FSF being slow like that!  I wasn't going to bother patching the
who program with this guys patch--redhat's rpm can't be more than a week
away....)

When the updated rpm did come out in July (!), I was shocked at the
"very hard to duplicate, thus the time to release the fix" part
of the comment.  So not only did redhat not accept a patch from
months earlier, they never commented publicly on any deficiencies
in the patch, or let anyone know they were having problems duplicating
this "elusive" bug.

Had they actually told any of us, I know I for one would have been
happy to help them; I am sure others would have as well.

I was and am slightly upset that they caused themselves to waste time
on the problem when they could have easily solved it quicker and more
efficiently by being public about their known problems.

(It would be nice to have a known-bugs tracking system accessable from
redhat's website that we could add our own annotations to.)

 -Mark Shewmaker
  mark at primefactor.com