Re: SUMMARY: "find" in root crontab hangs Solaris 2.2

From: David Allan Steffens (
Date: Tue Nov 23 1993 - 13:48:21 CST

In article <2ciuq6$> (Larry Broda) writes:
> Summary: a bug in hsfs is the cuplrit

Right so far. And this can be _cured_ by installing patch 101028...
        Patch-ID# 101028-01
        Synopsis: SunOS 5.2: system may hang due to memory leak in hsfs
        Date: Jun/04/93

>All responses agreed on the cause...

Unfortunately, none of your correspondents got it exactly right. I'm
sorry that I didn't manage to respond in time, but since I just posted
on this topic a week or two ago, I didn't think it was necessary.
However, now it seems that additional clarification is required.

>I will merely reproduce Bob Dowlings response below as it gives a
>comprehensive summary and solution for the problem:
>Do you have a CD inserted? If you do then this is bug 1133445,
>whose solution (from Sun) I post below. [ deleted to save bandwidth ]

Well, this is a good description of the _problem_ all right, and it is
also _a_ solution, but it is hardly _the_ solution, since it is
neither necessary nor sufficient. It is not _necessary_ since
installation of patch 101028 allows the system to run just fine with
the stock Solaris 2.2 crontab as mine has for the last several weeks.
It is also not _sufficient_ since the memory leak in the hsfs driver
is not fixed by munging crontab and you will eventually tickle this
bug some other way, e.g. by reading lots of AnswerBook entires, perhaps?
In fact, I'm almost _certain_ this hsfs memory bug bit me several
times in _exactly_ that way one afternoon. But I didn't recognize it
at the time since I was too busy breaking lots of other stuff. ;-)

>Sun have taken option 3 of the set below and used it in Solaris 2.3...

That is too bad. I've never been a fan of this particular crontab
entry ever since I first saw it (SunOS4-3.2?), and it is probably long
past time that it was corrected. However, the fact that Sun is
telling customers that fixing crontab fixes the hsfs hang problem in
Solaris 2.2 doesn't say much for the quality of Sun Software Support.
And I hope that patch 101028 was _also_ rolled into Solaris 2.3, because
otherwise there is still a pesky problem lurking in the wonderworkings.

For the sake of "completeness", I append my original posting. Note
that I sent this to Sun also. The surprising thing is that SRDB 5853
was _not_ in the SunSolve online databases at the time of my original
search, but appeared there shortly after my dialog with Sun hotline
support personnel. I hope that my suggestion to add a few additional
keywords and/or lines of problem description to the docs for patch
101028 was not transmogrified into SRDB 5853!

---------- begin included text ----------
Subject: nightly Solaris 2.2 hang can be cured by hsfs patch
Keywords: Solaris 2.2 SunOS 5.2 hsfs hang CDROM AnswerBook
Message-ID: <1456@eplunix.UUCP>
Date: 10 Nov 93 22:34:26 GMT

I recently sent the following note to the Sun bugs hotline.
I am posting it here because a few articles describing similar
problems have appeared in these NewsGroups over the past few
months, but I never saw a response with good solution for them.
Subject: suggestion for update to Solaris 2.2 patch documentation

Ever since I first installed it, I've been having an intermittant problem
with my Solaris 2.2 system on an LX. The system would be running
fine when I left work at the end of the day, but was completely
wedged when I returned to work the following morning. After this
happened a few times, I realized that it only occurred when I left
the Solaris 2.2 AnswerBook CD-ROM installed and mounted overnight.

After an exhausing search of the SunSolve online databases, I stumbled
across the following patch:
        Patch-ID# 101028-01
        Keywords: hsfs upgrade hangs
        Synopsis: SunOS 5.2: system may hang due to memory leak in hsfs
        Date: Jun/04/93

Although the documentation doesn't say so, this patch allows me to keep
the AnswerBook CD-ROM (or any other hsfs-format CD-ROM with many symbolic
links) online and mounted overnight. Without this patch, my system wedged
every night at 3AM when cron ran find down the CD-ROM directory tree.
Since I installed this patch, I've had the AnswerBook online and mounted
for over a week and haven't had a _single_ crash.

I'd like to suggest that you add a comment about this behavior to the
documentation for patch 101028. It might save someone else some time.
---------- end included text ----------

David Allan Steffens       | I believe in learning from past mistakes...
Eaton-Peabody Laboratory   | but why does a good education require so many?
Mass. Eye & Ear Infirmary, 243 Charles Street, Boston, MA 02114
{harvard,mit-eddie}!eplunix!das    Tel (617) 573-3748    Fax (617) 720-4408

This archive was generated by hypermail 2.1.2 : Fri Sep 28 2001 - 23:08:30 CDT