- This topic has 13 replies, 8 voices, and was last updated 16 years, 4 months ago by
.
-
Topic
-
We have installed CL 5.6 with Linux Redhat 5.0. When trying to run a cron job, the script cannot find CL_INSTALL_DIR . I know with 5.6, it’s replacing QUOVADX_INSTALL_DIR. Should the CL_INSTALL_DIR and FPATH be in the /etc/environment file like it is in AIX? Should the install script have put it there? Our does not have that there. In fact, the file exists but is empty. My guess is that I need to add both the CL_INSTALL_DIR and FPATH to that file and reboot the machine for it to take effect. I checked the documentation, but did not find any reference to that. Any help would be appreciated.
Thanks.
-
Replies
-
-
-
When you execute a command in cron on Linux the environment file .cshrc or .profile will not be used. So the environment variables are not set. To execute the environment file use bash –login prior to your command. Then the .profile will be executed and your environment should be set.
Greetings from germany,
Dirk
-
I have installed QDX 5.6 on AIX 5.3 and here are the lines that the install script added to my /etc/environment file Code:FPATH=/quovadx/qdx5.6/integrator/kshlib
CL_INSTALL_DIR=/quovadx/qdx5.6I did not have to reboot the server for them to take effect.
FYI – The install script needs to be run as root to have the necessary privilege to change files like /etc/environment.
Russ Ross
RussRoss318@gmail.com -
I’m going to hijack this thread a little…. I’m having the same issue w/AIX and the 5.6 install.
When I run my cron jobs (under hci) I don’t have the QUOVADX_INSTALL_DIR environment variable set.
I’ve tried adding the export in /hci/quovadx/qdx5.6/integrator/productinfo/defaults and in the .profile but neither of these work for me from my cron jobs.
I’m using a couple of techniques in the cron scripts e.g.:
exec $QUOVADX_INSTALL_DIR/integrator/bin/hcitcl “$0″ ${1+”$@”}
and
eval `/hci/quovadx/qdx5.3/integrator/sbin/hcisetenv -root ksh $2 $site`
that give me the following error:
QUOVADX_INSTALL_DIR is not set. Exit.
My question is way are your thoughts on the best way to correct this?
Should I change the code to point to CL_INSTALL_DIR or use another technique to set QUOVADX_INSTALL_DIR?
Thanks for any thoughts.
-
-
Is the default sh in AIX ksh? IIRC there is a way to start ksh as a login without being interactive. In a GNU/Linux system you’d run bash with the -l to do it, but
I can’t remember what (or if) it might be in ksh.
Of course, if I’d just scanned up a little I might have avoided this redundant post.
–kevin (
props 2 MH, heh) kinnell -
I placed both the CL_INSTALL_DIR and FPATH variables in the /etc/environment file, rebooted the machine, created a script just to echo them out, and placed them in cron. When it runs, it does not have any value for those variables. Who else runs Linux and what have you done to get cron to recognize (other than having the script have to log into an account). -
For cron jobs, I set up the environment like this: ~/.profile.cron
Code:CL_INSTALL_DIR=/hci/quovadx/qdx5.6
PATH=/usr/bin:/etc:/usr/sbin:/usr/ucb:$HOME/bin:/usr/bin/X11:/sbin:.:/usr/local/bin:/usr/local/lib:/usr/local/scripts:${CL_INSTALL_DIR}/integrato
export PATH
setrootAnd then in I run the scripts in cron like this:
Code:05 06,07,08,09,10,11,12,13,14,15,16,17,18,19,20,22 * * * . ~/.profile.cron;/hci/bin/monitor/edb_check.tcl sitebn
05 06,07,08,09,10,11,12,13,14,15,16,17,18,19,20,22 * * * . ~/.profile.cron;/hci/bin/monitor/edb_check.tcl sitebnoHere’s a portion of the script that demonstrates how to setup the environment in a tcl script:
Code:#!/hci/quovadx/qdx5.6/integrator/bin/tcl
global HciRoot
set debug 0
if { $argc != 1 } {
puts “Usage: $argv0 siteName”
exit
}set hostName [exec hostname]
if {$debug} {puts “hostName: $hostName”}
set hciSite [lindex $argv 0]
set date [clock format [clock seconds] -format %Y%m%d]# Set up the environment for Cloverleaf commands
eval [exec $HciRoot/sbin/hcisetenv -root tcl $HciRoot $hciSite]
set hciDir $HciRoot/$hciSite/scripts/errors# Get the number of lines in the error database. Ignore the header and footer (8 lines).
catch { exec hcidbdump -e | wc -l } lines
set lines [expr $lines – 8]
if {$debug} {puts “lines: $lines”}-- Max Drown (Infor)
-
We have set up the variables in the hcicrontab before we run any jobs (5.3 on RHE). Here’s an example
crontab -l: Code:FPATH=/quovadx/qdx5.3/integrator/kshlib
HCISITEDIR=/quovadx/qdx5.3/integrator/production
HCIROOT=/quovadx/qdx5.3/integrator
HCISITE=production
TCL_LIBRARY=/quovadx/qdx5.3/integrator/tcl/lib/tcl8.3
TCLX_LIBRARY=/quovadx/qdx5.3/integrator/tcl/lib/tclX8.2
QUOVADX_INSTALL_DIR=/quovadx/qdx5.3
LD_LIBRARY_PATH=/quovadx/qdx5.3/integrator/clgui/java/lib/i386:/quovadx/qdx5.3/integrator/clgui/java/lib/i386/server…59 23 * * * /path_to/some_command.tcl
Ugly, but it works until you figure out a better way. You can add extra environment
vars to control what cron does too; in case you want to change the default
MAILTO for error reports, or whatever else.
-
-
-
-
Hi folks, as I wrote before, cron doesn’t read the environment file (or .profile, .bashrc) by default.
To easily use the standard environment file you’ll have to use the shell command before executing the command:
Linux: bash –login
HP: ksh -m or -l
Sun: tcsh -m
That let cron read the environment file first before executing the command. I’ve done this on serveral Linux systems and my cron jobs then work normally.
Dirk
-
- The forum ‘Cloverleaf’ is closed to new topics and replies.
