Re: [Condor-users] condor installation problems

[Abdul Subhan <subhan@xxxxxxxxxxxxxxxxx>]

Hi
For all those having trouble with condor installation, these steps can make life 
easier.

1.Download latest version of Condor(in .tar fomrmat) from www.cs.wisc.edu to 
/usr/local/src


2. Unpack Condor and cd into the Condor directory


3. Create a Condor user and group (adduser condor –c “Condor Daemon”)


4. Run condor_install


5. Press Enter as it says on the prompt, and then answer the following questions 
as suggested


6. Are you planning to setup Condor on multiple machines? yes


7. Enter each hostname (include Master node): Manager C1 C2


8. There will not be a Condor release directory if this is a new install. Put 
the release directory in: /home/condor/release


9. Enter an e-mail address to receive mail about condor: root


10. Choose a mail program (/bin/mail)


11. Do all machines in your pool share a common filesystem? yes


12. Do all users have a unique id? Yes


13. Are any users not listed in passwd file: No


14. Enable java universe support: No


15. Shall I create links in some other directory? no


16. What machine is central manager? Master node (Manager)


17. Is condor home directory shared? Yes


18. Create subdirectories? Yes


19. Local partition for lock files? Yes


20. Where? /var/lock/condor


21. Machine-specific config files in one directory? Yes


22. Accept default location of /home/condor/release/etc

23. Give the pool a name (HADEED)


24. Should I put in a soft link from /home/condor/condor_config to 
/home/condor/release/etc/condor_config? Yes


25. Run /home/condor/sbin/condor_init


26. Rerun condor_install


27. Re-enter the machine name (manager.pel.ccse.kfupm.edu.sa)


28. Full installation: No


29. Submit-only: No


30. Condor Central Manager: Yes


31. It will ask how many machines, but the only possible answer is 1 here


32. You'll have to tell it where you put the Condor release directory: 
/home/condor/release


33. Re-enter pool name (HADEED)


34. Run /home/condor/sbin/condor_master

The Steps Below are optional, and can be done after u are satisfied with the 
installation

35. Copy /home/condor/release/etc/examples/condor.boot to /etc/rc.d/init.
d/condor


36. Create symbolic links:


ln -s /etc/rc.d/init.d/condor /etc/rc.d/rc3.d/S88condor

ln -s /etc/rc.d/init.d/condor /etc/rc.d/rc4.d/S88condor

ln -s /etc/rc.d/init.d/condor /etc/rc.d/rc5.d/S88condor


37. Repeat steps 35 and 36 on the cluster nodes by doing the following:


cpush /home/condor/release/etc/examples/condor.boot /etc/rc.d/init.d/condor

cexec ln -s /etc/rc.d/init.d/condor /etc/rc.d/rc3.d/S88condor

cexec ln -s /etc/rc.d/init.d/condor /etc/rc.d/rc4.d/S88condor

cexec ln -s /etc/rc.d/init.d/condor /etc/rc.d/rc5.d/S88condor


38. Set Path on master and cluster nodes by adding the following line to 
/etc/profile:


pathmunge /home/condor/release/bin


39. Cpush /etc/profile to the nodes

Quoting "Sathish S. Vadhiyar" <vss@xxxxxxxxxxxxxxxxxx>:

> 
> Hi,
> 
> Iam trying to install condor rpm on Fedora Core 2. Due to conflicting
> statements in the installation document regarding "root" and user
> "condor", Iam highly confused.
> 
> 1. What is the group ID I need to give to user "condor"? Is it "users"?
> 
> 2. Somewhere, it is said that during rpm installation,
> "If the condor user is present on the system, the installation script
> will
> assume that that is the effective user that Condor should run as."
> How can we verify this?
> 
> 3. After, condor has been installed to /usr/local/condor, I copy the
> /etc/condor_config file to /home/condor
> and change few minor things:
> ##  Where is the local condor directory for each host?
> LOCAL_DIR               = $(TILDE)/hosts/$(HOSTNAME)
> 
> ##  Where is the machine-specific local config file for each host?
> LOCAL_CONFIG_FILE       = $(LOCAL_DIR)/condor_config.local
> 
> I also make that the appropriate directories and files exist in
> /home/condor. I also create directories
> /home/condor/hosts/garl-intel1/execute
> /home/condor/hosts/garl-intel1/log
> /home/condor/hosts/garl-intel1/spool
> 
> where garl-intel1 is the $(HOSTNAME)
> 
> 4. And when I run the master as "root", I get:
> >> /usr/local/condor/sbin/condor_master
> Can't open "/home/condor/hosts/garl-intel1/log/MasterLog"
> 9/1 22:08:25 dprintf() had a fatal error in pid 1222
> Can't open "/home/condor/hosts/garl-intel1/log/MasterLog"
> errno: 13 (Permission denied)
> euid: 1, ruid: 0
> 
> The permissions of the directories in /home/condor/hosts/garl-intel1 are
> as follows:
> 
> -rw-r--r--  1 condor condor  585 Sep  1 22:05 condor_config.local
> drwxr-xr-x  2 condor condor 8192 Sep  1 22:07 execute
> drwxr-xr-x  2 condor condor 8192 Sep  1 22:07 log
> drwxr-xr-x  2 condor condor 8192 Sep  1 22:07 spool
> 
> 
> Now Iam completely lost. I have been trying different things for
> successful and correct installation and starting of daemons for the past
> several hours. What am I doing wrong? Please help me !
> 
> Thanks,
> Sathish
> 
> _______________________________________________
> Condor-users mailing list
> Condor-users@xxxxxxxxxxx
> http://lists.cs.wisc.edu/mailman/listinfo/condor-users
> 
> 



ABDUL SUBHAN

RESEARCH ASSISTANT 
COMPUTER ENGINEERING DEPARTMENT
 
P.O BOX # 7851 
KING FAHD UNIVERSITY OF PETROLEUM & MINERALS
DHAHRAN, 31261 
KINGDOM OF SAUDI ARABIA 
PHONE RESI: 00966-3-860-8000-EXT: 9902126