SUMMARY: df does'nt work anymore after a CLU_UPGRADE to 5.1A

From: <emanuele.lombardi_at_casaccia.enea.it>
Date: Mon, 04 Feb 2002 09:22:47 +0100 (CET)

Dear friends,
at last everithing is fine!

This summary is related to 2 mails I sent in October.

I installed the first patch for 5.1A
        t64v51ab1-c0008900-12954-e-20020124.tar
and now there are no more problem with AUTOFS and DF

Thanks to Tom Blinn and to the Engeneering Team.


Emanuele



My 1st mail was the following:
======================================================================

Hi TruFriends,

I'm a so-much-Tru64-fanatic that I really could'nt wait: as soon as I
received the 5.1A distribution I upgraded my 2 machine TruCluster 5.1 !

Everithing went smooth and easily as documented.

BUT now I have troubles with df (/usr/bin/df)

It does'nt work witouth arguments.
It does'nt work giving the domain#fset
It works giving the mounts points
It works giving -t advfs

When df does'nt work it simply stops waiting forever

root 599771 0.0 0.0 2.41M 312K pts/11 I + 17:05:00 0:00.01 df marte#m3
root 599165 0.0 0.0 2.38M 344K ?? I 17:01:42 0:00.04 /sbin/df


This behavour has conseguences upon setld (for istance), infact I was
trying to delete the old WEBES kit but setld hanged in the following
way:

-+- 591460 root setld -d WEBESBASE310
 \-+- 591824 root //usr/.smdb./WEBESBASE310.scp
   \-+- 591907 root mkcdsl -a /usr/opt/compaq/svctools/saved_profiles
     \-+- 591915 root sed -n -e s/.*% *\(\/..*\)/\1/p
       \--- 591913 root df


I'll be glad to everybody helping me!
Thank you very much from Italy,

Emanuele




My 2nd mail was the following:
==================================================================

  TruCluster 5.1A (freshly upgraded from 5.1 3th patch) is made of two
  member (an ES40 and a AS4100) and it worked properly until the upgrade
  process (which has been succesful and without any problem).

Dear TruFriends,

Thanks to the many suggestions coming from the list,
I now know that the problem arises from AUTOFS
Infact as autofsd starts it tries to mount ALL and EACH of the files
listed in /etc/auto.direct
Apart that it is not the supposed behavour, the problem arises when any of
the serving machine are down (as in my case of course!) and df returns its
info with a very long delay (while mount without arguments replies soon).

 Here my /etc/auto.master
/- /etc/auto.direct -rw,intr,soft,bg

 and here my /etc/auto.direct
/tmp_mnt/MS -rw,intr,soft masaccio:/MS_a/users
/tmp_mnt/CV -rw,intr,soft caravaggio:/CV_a/users

The situation is as follows:

/usr/bin/df given without arguments or with a domain#fset

still works badly (hangs and wait for a very long time). It happens
reguardles the value of the CMD_ENV environment variable.

df DOES'NT WORK
df marte#f1 DOES'NT WORK
df /dev/disk/cdrom3c DOES'NT WORK

The surprising thing is that giving df any fstype or file system let df
work properly!

df /var WORKS
df -t nfs WORKS
df -t advfs WORKS
df -t cdrom WORKS
df -t procfs WORKS
df -n WORKS

I noticed the malfunction before the "clu_upgrade clean"
and I cheched that even the old /usr/bin/.Old..df worked badly on the
upgarded system.

The whole cluster is working properly without any other (up to now
noticed!) problem.

Not even to say I can't open a software call since it is more than 3
months I bought the systems!

Thank you again from Italy,
Emanuele




-- 
$$$ Emanuele Lombardi
$$$ mail:  AMB-GEM-CLIM ENEA Casaccia
$$$        I-00060 S.M. di Galeria (RM)  ITALY
$$$ mailto:emanuele.lombardi_at_casaccia.enea.it
$$$ tel	+39 06 30483366 fax	+39 06 30484264
$$$
$$$                                |||
$$$                                \|/  ;_;
$$$ What does a process need        |   /"\
$$$ to become a daemon ?            |   \v/
$$$                                 |    | 
$$$ - a fork                        o---/!\---
$$$                                 |   |_|
$$$                                 |  _/ \_
$$$* Contrary to popular belief, UNIX is user friendly.
$$$  It's just very particular about who it makes friends with.
$$$* Computers are not intelligent, but they think they are. 
$$$* True programmers never die, they just branch to an odd address
$$$* THIS TRANSMISSION WAS MADE POSSIBLE BY 100% RECYCLED ELECTRONS
Received on Mon Feb 04 2002 - 08:23:26 NZDT

This archive was generated by hypermail 2.4.0 : Wed Nov 08 2023 - 11:53:43 NZDT