URGENT: Check Disk quotas

[Ahmed M. AlShafiy]

when i try to connect to oracle application user or oracle user it give me this message:

please checking for disk quotas

[Ahmed M. AlShafiy]

i try to switch user under root user to the applmgr user but it go to the next line and there is no response.


and when i try to connect thru reflation X
i get this message:

Last successful login for applmgr: Sun Apr 2 10:38:27 MST7MDT 2006 on pts/7
Last unsuccessful login for applmgr: Mon Feb 6 12:09:33 MST7MDT 2006 on pts/1
Please wait...checking for disk quotas

[Greg Vaidman]

the message you see is normal. if you never get back to a shell prompt, it's likely something else in your profile that is causing the problem. check the .profile of the user who is having this problem.

[john korterman]

Hi,

There could be many reasons for this behaviour, for instance that the max. limit of processes for a particular user is reached.
Try checking the number of active processes for the oracle user:
# ps -ef | grep oracle| wc -l
and compare that to the kernel parameter maxuprc.

regards,
John K.

[Ahmed M. AlShafiy]

this what i have:

# ps -ef | grep oracle|wc -l
2
# ps -ef | grep applmgr|wc -l
3


what is the way to kill it safe or foce it to be kill coz i try thru SAM but don not kill it.

[john korterman]

Hi,

judging from your feedback the problem is definitely not that the system has reached the max limit for user processes; hence there is no need for killing any oracle processes.
My guess was wrong.

I suggest two other apporaches:
Investigate what causes the login sequence for oracle to hang. If oracle is a sh/ksh/posix shell user, try this as root:
# sh -x ~oracle/.profile
and see how far you get: I assume that you must get stuck somewhere.

Check also /var/adm/syslog/syslog.log and root's mail for unusual messages.

regards,
John K.

[Ahmed M. AlShafiy]

when i run the command:
#sh -x ~oracle/.profile ==> i get ==>

+ eval export TERM; TERM=xterm;
+ export TERM
+ TERM=xterm
+ stty erase ^H kill ^U intr ^C eof ^D
+ stty hupcl ixon ixoff
+ tabs
+ set -u
+ trap echo 'logout' 0
+ EDITOR=vi
+ export EDITOR
+ umask 022
+ ORACLE_TERM=vt220
+ export ORACLE_TERM
+ . /oranas/oracle/nasdb/8.1.7/NAS.env
+ ORACLE_HOME=/oranas/oracle/nasdb/8.1.7
+ export ORACLE_HOME
+ PATH=/oranas/oracle/nasdb/8.1.7/bin:/usr/ccs/bin:/usr/sbin:/usr/sbin:/usr/bin:/usr/ccs/bin:/usr/contrib/bin:/opt/hparray/bin:/opt/nettladm/bin:/opt/upgrade/bin:/opt/fcms/bin:/opt/pd/bin:/opt/resmon/bin:/usr/bin/X11:/usr/contrib/bin/X11:/opt/graphics/common/bin:/opt/gnome/bin:/opt/mozilla:/opt/mx/bin:/usr/sbin/diag/contrib:/opt/wbem/bin:/opt/wbem/sbin:/opt/ignite/bin:/opt/aCC/bin:/opt/langtools/bin:/sbin:/home/root
+ LD_LIBRARY_PATH=/oranas/oracle/nasdb/8.1.7/lib:/usr/dt/lib:/usr/openwin/lib
+ LIBPATH=/oranas/oracle/nasdb/8.1.7/lib:/usr/dt/lib:/usr/openwin/lib
+ LINK_CNTRL=L_PTHREADS_D7
+ SHLIB_PATH=/oranas/oracle/nasdb/8.1.7/lib:/usr/lib
+ TNS_ADMIN=/oranas/oracle/nasdb/8.1.7/network/admin/NAS
+ export PATH LD_LIBRARY_PATH SHLIB_PATH LIBPATH LINK_CNTRL TNS_ADMIN
+ unset TWO_TASK
+ unset ORA_NLS33
+ ORACLE_SID=NAS
+ export ORACLE_SID
+ NLS_LANG=American_America.AR8MSWIN1256
+ export NLS_LANG
+ NLS_DATE_FORMAT=DD-MON-RR
+ export NLS_DATE_FORMAT
+ NLS_NUMERIC_CHARACTERS=.,
+ export NLS_NUMERIC_CHARACTERS
+ NLS_SORT=binary
+ export NLS_SORT
+ echo logout
logout
#

[john korterman]

Hi again,

sorry, I have not been as attentive as I should have been!
The problem concerns of course not the oracle user, but the applmgr user.
If his shell is the same as for the oracle user, try this check:
# sh -x ~applmgr/.profile

which then must hang somewhere?!

regards,
John K.

[Bill Hassell]

When this message occurs just prior to a hang, it is quite often a bad filesystem, most likely an NFS mounted filesystem from another computer. The same hang will usually occur with bdf too. Does bdf hang? Do you hsve NFS mounts?

[Ahmed M. AlShafiy]

it dose not hang and the result for the command:

#sh -x ~applmgr/.profile

+ test /oranas/applmgr/nascomn =
+ test a != s
+ test ! -d /oranas/applmgr/nascomn
+ test NAS.env = adsetenv.sh
+ + echo OAM_TOP:s
+ awk -F: {print $1}
adov_var_name=OAM_TOP
+ + echo OAM_TOP:s
+ awk -F: {print $2}
adov_check_type=s
+ eval adov_var_value=$OAM_TOP
+ adov_var_value=/oranas/applmgr/nascomn/java/oracle/apps
+ test /oranas/applmgr/nascomn/java/oracle/apps =
+ test s != s
+ + echo CLASSPATH:s
+ awk -F: {print $1}
adov_var_name=CLASSPATH
+ + echo CLASSPATH:s
+ awk -F: {print $2}
adov_check_type=s
+ eval adov_var_value=$CLASSPATH
+ adov_var_value=/oranas/applmgr/nascomn/util/jre/1.1.8/lib/rt.jar:/oranas/applmgr/nascomn/util/jre/1.1.8/lib/i18n.jar:/oranas/applmgr/nascomn/java/appsborg.zip:/oranas/applmgr/nascomn/java/apps.zip:/oranas/applmgr/nasora/8.0.6/forms60/java:/oranas/applmgr/nascomn/java
+ test /oranas/applmgr/nascomn/util/jre/1.1.8/lib/rt.jar:/oranas/applmgr/nascomn/util/jre/1.1.8/lib/i18n.jar:/oranas/applmgr/nascomn/java/appsborg.zip:/oranas/applmgr/nascomn/java/apps.zip:/oranas/applmgr/nasora/8.0.6/forms60/java:/oranas/applmgr/nascomn/java =
+ test s != s
+ + echo LD_LIBRARY_PATH:s
+ awk -F: {print $1}
adov_var_name=LD_LIBRARY_PATH
+ + echo LD_LIBRARY_PATH:s
+ awk -F: {print $2}
adov_check_type=s
+ eval adov_var_value=$LD_LIBRARY_PATH
+ adov_var_value=/oranas/applmgr/nasora/8.0.6/network/jre11/lib/PA_RISC/native_threads::/oranas/applmgr/nasora/8.0.6/lib:/usr/lib
+ test /oranas/applmgr/nasora/8.0.6/network/jre11/lib/PA_RISC/native_threads::/oranas/applmgr/nasora/8.0.6/lib:/usr/lib =
+ test s != s
+ test No = Yes
+ test HP9000S800 != M88OPEN
+ adov_unset_it=unset
+ eval unset adov_check_vars
+ unset adov_check_vars
+ eval unset adov_bad
+ unset adov_bad
+ eval unset adov_loop_index
+ unset adov_loop_index
+ eval unset adov_var_name
+ unset adov_var_name
+ eval unset adov_check_type
+ unset adov_check_type
+ eval unset adov_var_value
+ unset adov_var_value
+ eval unset oa_htmlbin
+ unset oa_htmlbin
+ test HP9000S800 != M88OPEN
+ unset adov_unset_it
+ FORMS60_WEB_CONFIG_FILE=/oranas/applmgr/nascomn/html/bin/appsweb.cfg
+ export FORMS60_WEB_CONFIG_FILE
+ REPORTS60_PRE=&5555
+ export REPORTS60_PRE
+ REPORTS60_POST=&5556
+ export REPORTS60_POST
+ export PERL5LIB=/oranas/applmgr/nasora/iAS/Apache/perl/lib/5.00503/PA-RISC2.0:/oranas/applmgr/nasora/iAS/Apache/perl/lib/5.00503
+ echo logout
logout

[Ahmed M. AlShafiy]

but when i try to restart the server it login to the user that i have try to login to them before.

[Ahmed M. AlShafiy]

i check the mail file under syslog and i find this:

Apr 1 00:01:07 oracle sendmail[7360]: gethostbyaddr(10.3.242.25) failed: 1
Apr 1 00:01:07 oracle sendmail[7360]: AAA07360: from=applmgr, size=209, class=0, pri=30209, nrcpts=1, msgid=<200604010701.AAA07360@nasjet1.nasaviation.com>, relay=applmgr@localhost
Apr 1 00:01:07 oracle sendmail[7362]: AAA07360: to=aalshafiy@nasaviation.com, ctladdr=applmgr (106/103), delay=00:00:00, xdelay=00:00:00, mailer=relay, relay=mail1.nasaviation.com. [10.3.242.40], stat=Sent (2.6.0 <200604010701.AAA07360@nasjet1.nasaviation.com> Queued mail for delivery)
Apr 1 00:02:40 oracle sendmail[7405]: gethostbyaddr(10.3.242.25) failed: 1
Apr 1 00:02:40 oracle sendmail[7405]: AAA07405: to=DBA, delay=00:00:00, mailer=local, stat=User unknown
Apr 1 00:02:40 oracle sendmail[7405]: AAA07405: from=applmgr, size=125, class=0, pri=30125, nrcpts=1, msgid=<200604010702.AAA07405@nasjet1.nasaviation.com>, relay=applmgr@localhost
Apr 2 14:40:06 oracle /sbin/init.d/sendmail[947]: #### rebooted ####
Apr 2 14:40:06 oracle sendmail[955]: gethostbyaddr(10.3.242.25) failed: 1
Apr 2 14:40:06 oracle sendmail[955]: alias database /etc/mail/aliases rebuilt by root
Apr 2 14:40:06 oracle sendmail[955]: NOQUEUE: SYSERR(root): /etc/mail/aliases: line 15: missing colon
Apr 2 14:40:06 oracle sendmail[955]: NOQUEUE: SYSERR(root): /etc/mail/aliases: line 18: missing colon
Apr 2 14:40:06 oracle sendmail[955]: NOQUEUE: SYSERR(root): /etc/mail/aliases: line 21: missing colon
Apr 2 14:40:06 oracle sendmail[955]: NOQUEUE: SYSERR(root): /etc/mail/aliases: line 24: missing colon
Apr 2 14:40:06 oracle sendmail[955]: NOQUEUE: SYSERR(root): /etc/mail/aliases: line 29: missing colon
Apr 2 14:40:06 oracle sendmail[955]: NOQUEUE: SYSERR(root): /etc/mail/aliases: line 32: missing colon
Apr 2 14:40:06 oracle sendmail[955]: /etc/mail/aliases: 11 aliases, longest 26 bytes, 165 bytes total
Apr 2 14:40:07 oracle sendmail[956]: gethostbyaddr(10.3.242.25) failed: 1
Apr 2 14:40:07 oracle sendmail[957]: starting daemon (8.9.3 (PHNE_29774)): SMTP+queueing@00:30:00

and i'm uploadfing the syslog

[john korterman]

Hi again,

since the .profile files can be sourced without any problems, they probably do not cause the login process to hang.
My understanding is that
# su - oracle
causes the login process to halt and the last message displayed is "Check Disk quotas".
Is that correct?

Does su to other users behave likewise?

regards,
John K.

[Ahmed M. AlShafiy]

i have tow users oracle and applmgr and when i type:

#su oracle or applmgr

the cursore go to the next line without any response and i left it for long time without any response and i try to connect thru a program called Reflection X it give me this message without connect:


Last successful login for applmgr: Mon Apr 3 16:57:01 MST7MDT 2006 on pts/3
Last unsuccessful login for applmgr: Mon Feb 6 12:09:33 MST7MDT 2006 on pts/1
Please wait...checking for disk quotas




please can you help me.

[OldSchool]

As Bill H noted above, this appears to be a problem w/ an nfs mount somewhere.

Does this happen for every user?

Try logging in to the system console as root. once you get logged in, whereever that may be, run a "bdf" command and see if it hangs or stalls.

examine the contents of /etc/mnttab for nfs mounts. try pinging the servers.

I believe the most likely scenario is that the you have one (or more) nfs mounts, and their is a connectivity issue between the nfs server and client.

[Ahmed M. AlShafiy]

# bdf
Filesystem kbytes used avail %used Mounted on
/dev/vg00/lvol3 4194304 2263912 1915336 54% /
/dev/vg00/lvol1 298928 69864 199168 26% /stand
/dev/vg00/lvol8 4825088 625536 4167296 13% /var
/dev/vg00/lvol7 6291456 1116400 5134632 18% /usr
/dev/vg00/lvol4 4194304 495344 3670096 12% /tmp
/dev/vg01/lvol1 81920000 53945544 27757112 66% /oranas
/dev/vg00/lvol6 3145728 1263288 1867744 40% /opt
/dev/vg00/lvol5 1048576 2624 1037840 0% /home
#


this is the result of bdf command

[Ahmed M. AlShafiy]

this is what i found under /etc/mnttab


/dev/vg00/lvol3 / vxfs log 0 1 1144010364
/dev/vg00/lvol1 /stand hfs defaults 0 0 1144010365
/dev/vg00/lvol8 /var vxfs delaylog,nodatainlog 0 0 1144010368
/dev/vg00/lvol7 /usr vxfs delaylog,nodatainlog 0 0 1144010368
/dev/vg00/lvol4 /tmp vxfs delaylog,nodatainlog 0 0 1144010368
/dev/vg01/lvol1 /oranas vxfs delaylog,nodatainlog,nolargefiles 0 0 1144010368
/dev/vg00/lvol6 /opt vxfs delaylog,nodatainlog 0 0 1144010368
/dev/vg00/lvol5 /home vxfs delaylog,nodatainlog 0 0 1144010368

[john korterman]

Hi again,

perhaps the users oracle and applmgr do not use the .profile files during startup. Please post the output of:
# grep oracle /etc/passwd
and
# grep applmgr /etc/passwd


regards,
John K.

[Ahmed M. AlShafiy]

the results:
# grep oracle /etc/passwd
oracle: *:107:103:oracle dba,,,:/oranas/oracle:/usr/bin/sh

# grep applmgr /etc/passwd
applmgr: *:106:103:Application User,,,:/oranas/applmgr:/usr/bin/sh

[john korterman]

Hi again,

Hmmmm.....
both users apparently make use of the .profiles that could be sourced without any problems. And the users' home dirs are also mounted.
The syslog errors concerning mail problems can hardly be related to this.
Does the problem still persist?
I must admit that I can offer no other explanation than perhaps a display problem for your screen (far out).
Does the system react differently for other users?
Perhaps other members have better ideas.

Regards,
John K.

[Bill Hassell]

Try using plain old telnet from your PC to bypass all the Xwindows code. When you use su with a - you are performing a complete login including profiles. Without the - the login bypasses the profiles.

I suspect there may be some special code in the login profiles, so put this in the front of /etc/profile:

set -x

This will trace each step as you login. The previous traces were run on .profile independent of the login process so this will at least isolate the step where the hang occurs.

[Ahmed M. AlShafiy]

can you please provide me with the full command?

[Bill Hassell]

Use vi to edit the file /etc/profile
Insert one line that looks like this:

set -x

Close the file. Now everyone that logs into your system will have all the steps in /etc/profile displayed. If there is a hyang, you will see the steps preceeding the hang. Note that you will have to remove the set -x line once you determine where the hang occurs.