deploy explorer via uce?
I'm trying to deploy explorer (SUNWexplu and SUNWexplo packages) via uce.My simulation mode works fine, but when I try to deploy, I get:
Job Name: PKG_INSTALL_JOB_SUNWexplu-5.4-2006.04.18.23.14
Job ID:34
Job Exec Line: /usr/sbin//pkgadd pkgadd -n -d /opt/local/aduva/director_agent//config_files/.Job_315/SUNWexplu -a /opt/local/aduva/director_agent//config_files/.Job_315/admin_file.21989_PLaO8Q all
Job Output:
ABOUT TO EXECUTE..../usr/sbin//pkgadd
Copyright 1996-2006 Sun Microsystems, Inc.
All rights reserved.
/var/lib/aduva/director_agent/config_files/.Job_315/SUNWexplu/SUNWexplu/install /checkinstall: /var/lib/aduva/director_agent/config_files/.Job_315/SUNWexplu/SUNWexplu/install /checkinstall: cannot open
pkgadd: ERROR: checkinstall script did not complete successfully
Installation of <SUNWexplu> failed.
No changes were made to the system.
Job FINISHED with Value: 1
*****
***************************************************************************
ERROR. Got Failure while running job: PKG_INSTALL_JOB_SUNWexplu-5.4-2006.04.18.23.14.
Am I doing something wrong, or can you describe the steps I need to take to successfully perform this action?
Thanks.
[1282 byte] By [
dougsc] at [2007-11-26 7:51:59]
# 1
Doug,We were able to initially re-create the same problem you explained but after attempting a manual install of the package and then removing it again we were able to successfully add the package through UCE.Can you verify what the UMASK is on your system?
# 2
umask is 0022:# umask0022# /bin/cshita-sb1000a# umask22ita-sb1000a#
# 3
The UMASK looks fine but it is generally a permissions issue that causes the problem when gettiong the cannot open checkinstall error.
Can you post the permissions of the directory tree down to the .JobXXX directory (i.e the permissions of the /var/lib/aduva/director_agent/config_files/ tree)
Also have you had problems running any other jobs on this host or is it just the explorer install that fails.
Finally, can you verify that disk space isnt an issue on the host (i doubt it is but just to be sure)
# 4
ita-sb1000a# pwd
/var/lib/aduva
ita-sb1000a# find . -type d -exec ls -ld {} \;
drwxr-xr-x3 rootother512 May 17 10:10 .
drwxr-xr-x8 rootother512 May 17 10:10 ./director_agent
drwxr-xr-x2 rootother512 May 17 10:10 ./director_agent/bin
drwxr-xr-x2 rootother4608 Jun 9 22:42 ./director_agent/blobs
drwxr-xr-x2 rootother1024 May 17 14:11 ./director_agent/blobs.save
drwxr-xr-x2 rootother1024 Jun 11 14:13 ./director_agent/config_files
drwxr-xr-x3 rootother1024 Jun 11 14:13 ./director_agent/logs
drwxr-xr-x2 rootother512 May 17 12:09 ./director_agent/logs/ita-sb1000a_job_96
drwxr-xr-x2 rootother512 May 17 14:29 ./director_agent/notifications
Also, I tried pkginstalling SUNWexplo & SUNWexplu manually, and that worked fine. Then I pkgrm'd them and tried the UCE job again and it failed in the same way it did last time.
# 5
ita-sb1000a# df -kl
Filesystemkbytesusedavail capacity Mounted on
/dev/dsk/c1t1d0s031984951 878447 307866553%/
/proc 0000%/proc
fd 0000%/dev/fd
mnttab 0000%/etc/mnttab
/dev/dsk/c1t1d0s5962573 349149 55567039%/var
swap360215216 36021361%/var/run
swap36056403504 36021361%/tmp
Also, I think I have deployed other jobs ok on this host before. I will try the explorer one on some other hosts in the morning.
# 6
So, I was able to deploy a different, patch-related job on ita-sb1000a successfully.Which would seem to rule out any umask or disk space issues. And I was unable to deploy the explorer job on another Solaris 8 host. Will try some S9/10 hosts, see if I get similar errors.
# 7
So, I was able to deploy explorer successfully on a freshly installed S10 host. Trying the same thing that I did on the S10 host on the S8 host failed just like it has been...
# 8
I have now created a multi-distro profile with explorer 5.4. I can run the profile successfully on S9 and S10 hosts to install explorer via uce. However, every S8 host I try fails in the same way as documented in this thread.Any ideas?
# 9
We will try and reproduce this problem our test lab and let you know what we find.
# 10
Unfortunately we have not been able to investigate this issue on a Solaris 8 machine.
However if you can enable debug mode and send us the logs it should hopefully help point to the problem.
To enable debug mode on the agent add the line:
( all ) ( invisible.debug.__group1.debug_mode, true );
to the "/opt/local/aduva/director_agent/bin/.director.rc" file on the Sol 8 system.
Then restart the agent, and try deploying the job again:
$ /etc/init.d/director_agent restart
Send us the contents of the "/opt/local/aduva/director_agent/logs/" directory for our analysis. You can email it to "rsc-forum-inbound@sun.com". Please use the following as the subject line of the email.
"Sun Forums Update Notice: Sun Update Connection Enterprise Evaluation: "deploy explorer via uce?""
# 11
Just to close this out, this is resolved in SC 1.1, and explorer deployment works for S8 as well as S9 and S10/S10x86.
