How to create user in Rocks

Creating user in Rocks is simple. And also because of the availability of lots of scripting tools build into the operating system made exporting username and password to client nodes is much more simpler than before.

Before I use Rocks in my cluster, I had used Red Hat 7.0, 7.3, 8.0, 8.1, 9.0, Fedora Core, FC2 and Slackware 9, 10, and 11 in my cluster. Except Slackware 11 where I'm tinkering with LDAP to do some authentication, I used NIS (Network Information Services) to manage the authentication at client nodes. At those time, I'm not that good in creating script using BASH or CSH (still not good today). But I managed to create some script to add user based on input file.

So, back to the topic. How to create user(s) in Rocks? Simple.

1. Use useradd command to a new user.
$ useradd myUser

2. Create password for the user
$ passwd myUser
Changing password for user myUser
New UNIX password :
Retype new UNIX password :
passwd : all authentication tokens updated successfully

After you sending the particular to the user, remind the user to change the password. It is a good practice if you can integrate some sort of tools to test the strong'ness' of user's password.

3. Use command rocks-user-sync to send the user's details throughout the cluster.
$ rocks-user-sync

There are many output will be displayed to your screen. Don't panic. It's means the process work smoothly (unless it mentioned fail somewhere in the output)

4. Ask your user to try to login into the account. You can do the test by login into the account. Or if you create 100 or more accounts, you can pick random account to test. After you login into the account, test login into other nodes in the cluster by using ssh. Lets say I has a compute node name compute-0-0.

$ ssh compute-0-0

5. If the client didn't prompt password during login, its mean everything works fine. Or if its prompt password and you can successfully login without any problem, it also means everything works fine.

There are other test to run to make sure the user's accounts work fine.

1. NFS test
Copy any file into the user account in the main node (or any node). If the file appear in other nodes, so, NFS work fine. If not, contact your admin (or if you are the admin, please troubleshoot NFS) so, they can troubleshoot the NFS.

2. SGE test
note to self : please put some test for SGE

3. MPICH test

Rocks Cluster : Bad FTP response

When you are trying to install a compute node (its already pass the Kickstart process), if suddenly your compute node returning "Bad FTP Response". Please, don't try too hard trying to figure out where the heck is the ftp daemon in your server (like I do). I spend the entire day to figure out why it is still got the statement. Please do like this (please, don't go look at Rocks Cluster mailing list. Sometimes, they give you lots of things to do before you can solve it. Follow my simple and easier step.

Please see the access.log and error.log in your /var/log/httpd/ directory. Look into the recent ones. Use tail will do.

192.168.0.254 - - [15/Aug/2007:14:17:58 +0800] "GET /install/rocks-dist/lan/i386/RedHat/base/updates.img HTTP/1.0" 403 330 "-" "-"
192.168.0.254 - - [15/Aug/2007:14:17:59 +0800] "GET /install/rocks-dist/lan/i386/RedHat/base/updates.img HTTP/1.0" 403 330 "-" "-"
192.168.0.254 - - [15/Aug/2007:14:18:00 +0800] "GET /install/rocks-dist/lan/i386/RedHat/base/updates.img HTTP/1.0" 403 330 "-" "-"
192.168.0.254 - - [15/Aug/2007:14:18:01 +0800] "GET /install/rocks-dist/lan/i386/RedHat/base/updates.img HTTP/1.0" 403 330 "-" "-"
192.168.0.254 - - [15/Aug/2007:14:18:02 +0800] "GET /install/rocks-dist/lan/i386/RedHat/base/updates.img HTTP/1.0" 403 330 "-" "-"
192.168.0.254 - - [15/Aug/2007:14:18:03 +0800] "GET /install/rocks-dist/lan/i386/RedHat/base/updates.img HTTP/1.0" 403 330 "-" "-"
192.168.0.254 - - [15/Aug/2007:14:18:04 +0800] "GET /install/rocks-dist/lan/i386/RedHat/base/updates.img HTTP/1.0" 403 330 "-" "-"
192.168.0.254 - - [15/Aug/2007:14:18:05 +0800] "GET /install/rocks-dist/lan/i386/RedHat/base/updates.img HTTP/1.0" 403 330 "-" "-"
192.168.0.254 - - [15/Aug/2007:14:18:06 +0800] "GET /install/rocks-dist/lan/i386/RedHat/base/updates.img HTTP/1.0" 403 330 "-" "-"
192.168.0.254 - - [15/Aug/2007:14:18:16 +0800] "GET /install/rocks-dist/lan/i386/RedHat/base/updates.img HTTP/1.0" 403 330 "-" "-"
192.168.0.254 - - [15/Aug/2007:14:18:17 +0800] "GET /install/rocks-dist/lan/i386/RedHat/base/updates.img HTTP/1.0" 403 330 "-" "-"
192.168.0.254 - - [15/Aug/2007:14:18:18 +0800] "GET /install/rocks-dist/lan/i386/RedHat/base/updates.img HTTP/1.0" 403 330 "-" "-"
192.168.0.254 - - [15/Aug/2007:14:18:19 +0800] "GET /install/rocks-dist/lan/i386/RedHat/base/updates.img HTTP/1.0" 403 330 "-" "-"
192.168.0.254 - - [15/Aug/2007:14:18:20 +0800] "GET /install/rocks-dist/lan/i386/RedHat/base/updates.img HTTP/1.0" 403 330 "-" "-"
192.168.0.254 - - [15/Aug/2007:14:18:21 +0800] "GET /install/rocks-dist/lan/i386/RedHat/base/updates.img HTTP/1.0" 403 330 "-" "-"
192.168.0.254 - - [15/Aug/2007:14:18:22 +0800] "GET /install/rocks-dist/lan/i386/RedHat/base/updates.img HTTP/1.0" 403 330 "-" "-"
192.168.0.254 - - [15/Aug/2007:14:18:23 +0800] "GET /install/rocks-dist/lan/i386/RedHat/base/updates.img HTTP/1.0" 403 330 "-" "-"
192.168.0.254 - - [15/Aug/2007:14:18:24 +0800] "GET /install/rocks-dist/lan/i386/RedHat/base/updates.img HTTP/1.0" 403 330 "-" "-"

In the access.log you will see that your node already get the file. But do it? Let's see the error.log.

[Wed Aug 15 14:17:58 2007] [error] [client 192.168.0.254] client denied by server configuration: /var/www/html/install/rocks-dist/lan/i386/RedHat/base/updates.img
[Wed Aug 15 14:17:59 2007] [error] [client 192.168.0.254] client denied by server configuration: /var/www/html/install/rocks-dist/lan/i386/RedHat/base/updates.img
[Wed Aug 15 14:18:00 2007] [error] [client 192.168.0.254] client denied by server configuration: /var/www/html/install/rocks-dist/lan/i386/RedHat/base/updates.img
[Wed Aug 15 14:18:01 2007] [error] [client 192.168.0.254] client denied by server configuration: /var/www/html/install/rocks-dist/lan/i386/RedHat/base/updates.img
[Wed Aug 15 14:18:02 2007] [error] [client 192.168.0.254] client denied by server configuration: /var/www/html/install/rocks-dist/lan/i386/RedHat/base/updates.img
[Wed Aug 15 14:18:03 2007] [error] [client 192.168.0.254] client denied by server configuration: /var/www/html/install/rocks-dist/lan/i386/RedHat/base/updates.img
[Wed Aug 15 14:18:04 2007] [error] [client 192.168.0.254] client denied by server configuration: /var/www/html/install/rocks-dist/lan/i386/RedHat/base/updates.img
[Wed Aug 15 14:18:05 2007] [error] [client 192.168.0.254] client denied by server configuration: /var/www/html/install/rocks-dist/lan/i386/RedHat/base/updates.img
[Wed Aug 15 14:18:06 2007] [error] [client 192.168.0.254] client denied by server configuration: /var/www/html/install/rocks-dist/lan/i386/RedHat/base/updates.img
[Wed Aug 15 14:18:16 2007] [error] [client 192.168.0.254] client denied by server configuration: /var/www/html/install/rocks-dist/lan/i386/RedHat/base/updates.img
[Wed Aug 15 14:18:17 2007] [error] [client 192.168.0.254] client denied by server configuration: /var/www/html/install/rocks-dist/lan/i386/RedHat/base/updates.img
[Wed Aug 15 14:18:18 2007] [error] [client 192.168.0.254] client denied by server configuration: /var/www/html/install/rocks-dist/lan/i386/RedHat/base/updates.img
[Wed Aug 15 14:18:19 2007] [error] [client 192.168.0.254] client denied by server configuration: /var/www/html/install/rocks-dist/lan/i386/RedHat/base/updates.img
[Wed Aug 15 14:18:20 2007] [error] [client 192.168.0.254] client denied by server configuration: /var/www/html/install/rocks-dist/lan/i386/RedHat/base/updates.img
[Wed Aug 15 14:18:21 2007] [error] [client 192.168.0.254] client denied by server configuration: /var/www/html/install/rocks-dist/lan/i386/RedHat/base/updates.img
[Wed Aug 15 14:18:22 2007] [error] [client 192.168.0.254] client denied by server configuration: /var/www/html/install/rocks-dist/lan/i386/RedHat/base/updates.img
[Wed Aug 15 14:18:23 2007] [error] [client 192.168.0.254] client denied by server configuration: /var/www/html/install/rocks-dist/lan/i386/RedHat/base/updates.img
[Wed Aug 15 14:18:24 2007] [error] [client 192.168.0.254] client denied by server configuration: /var/www/html/install/rocks-dist/lan/i386/RedHat/base/updates.img

The same date and time. See that the client is denied by the server. What's the heck is the configuration you may ask. Afraid not my young padawan. It is the httpd configuration. But is it? I spend 10 minutes reading httpd.conf until I read the last line.

include conf.d/rocks.conf

So, the httpd get another set of configuration from conf.d. So, lets see what are the files in conf.d.

-r--r--r-- 1 root root 361 Aug 15 21:51 411.conf
-r--r--r-- 1 root root 1417 Aug 15 21:52 central.conf
-r--r--r-- 1 root root 446 Aug 15 21:52 myadmin.conf
-rw-r--r-- 1 root root 448 Sep 21 2006 php.conf
drwxr-xr-x 2 root root 4096 Jun 27 09:07 RCS
-rw-r--r-- 1 root root 392 Aug 13 2006 README
-r--r--r-- 1 root root 774 Aug 15 21:50 rocks.conf
-r--r--r-- 1 root root 268 Jun 27 09:06 rocks-wordpress.conf
-rw-r--r-- 1 root root 10919 Aug 13 2006 ssl.conf
-rw-r--r-- 1 root root 299 Aug 13 2006 welcome.conf
-rw-r--r-- 1 root root 138 Sep 25 2006 wordpress.conf

You need to change central.conf, myadmin.conf and rocks.conf. Compare your private IP with the configuration file. Make sure the IP in the files is your private IP not your public IP. Try to retrieve the image again. I'm sure you will successfully retrieve the image.

Try it my young padawan. May the force be with you!

Rocks Cluster : Editing Network Configurations

After installing Rocks, and dirtying your hands playing with the configurations, then to your horror you realize that the setting you put for your eth0 and eth1 is not like you intended. Or you want to edit the IP address for eth0 or eth1 as given by your network administrator. So, what you can do?

In Rocks, the configuration is store in configuration files and mysql database. So, how to edit the network configuration?

First, login into your mysql console.

$ mysql -u root -p
Enter password :

Enter your root password for the machine. For example, if you use 'abcd1234' (don't use this password, it is a weakest password), put in 'abcd1234'.

You can view current databases available in your server.

mysql> show databases;
+-----------+
| Database |
+-----------+
| cluster |
| mysql |
| test |
| wordpress |
+-----------+
4 rows in set (0.00 sec)

You can ignore mysql, test and wordpress unless you want to do something on the 3 databases. For now, we will concentrate on cluster database.

mysql> use cluster;
Database changed

You can use show tables command on the console to list all the tables available in the cluster database. But for now, I will concentrate on network table.

mysql> select * from networks;
+----+------+-------------------+-------------+---------------+---------------+
| ID | Node | MAC | IP | Netmask | Gateway | Name | Device | Module | Options | Comment |
+----+------+-------------------+-------------+---------------+---------------+
| 1 | 1 | theMac1| 192.168.1.3 | 255.255.255.0 | 192.168.1.254 | yourhost.yourdomain | eth0 | 3c59x | NULL | NULL |
| 2 | 1 | theMac2 | 192.168.0.1 | 255.255.255.0 | NULL | yourhost | eth1 | tg3 | NULL | NULL |
+----+------+-------------------+-------------+---------------+---------------+
2 rows in set (0.00 sec)

As you can see from the list, most of the network configuration is in the database. So, let say if you want to change the IP address from local into public IP use by your organization. And lets say the device that will bear the changes fall into eth0. So, do this sql statement.

mysql> update networks set ip='into.your.org.ip' where device='eth0';

Exit from mysql. Then execute this statement.

$ dbreport ifcfg eth0 yourhostname >> /etc/sysconfig/network-scripts/ifcfg-eth0

Restart your network or reboot your system to make sure the changes took effect.

This article is a part to help me in building the cluster and prevent me from digging into the Rocks mailing list. ;)

Installing Rocks 4.2.1 - for frontend

This entry is to guide me on how to install Rocks into my cluster. It is for me only but you can read it also. I'm tired searching the Rocks mailing list so, I've decided to put the solution (that I found) here in this blog.

When at first I install Rocks, it never get pass to the X-Window (Graphical User Interface - GUI) of Rocks. The problem is like this :

Attempting to start native X Server
Waiting for X Server to start....log located in /tmp/X.log
1..2..3..4..5.. X Server started successfully
Starting graphical installation...
The application 'mini-wm' lost its conection to the display :1.0
Most likely X server was shut down or you killed/destroyed the application
Terminated
Sending termination signals ....done
Sending kill signals ...done
disabling swap...
unmounting filesystem...
/mt/runtime umount failed (16)
disabling /dev/loop2 LOOP_CLR_FD failed : 16
/proc/bus/usb done
/proc umount failed (16)
/dev/pts done
/sys done
/tmp/ramfs umount failed (16)
/mnt/cdrom done
/export done
rebooting system
kernel panic - not syncing : VFS : Unable to mount root fs on unknown block (8,3)

So, I search the mailing list for the problem. It has 2 solutions. The first solution, make sure your frontend (or master node) has 1 GB of RAM. Less than that, you may encounter the problem. The other solution is by doing this (I've got this from here) :

In the ks.cfg file in the Kernel CD, the following commands for Anaconda are given:

dispatch.skipStep("handleX11pkgs", permanent = 1)
dispatch.skipStep("videocard", permanent = 1)
dispatch.skipStep("monitor", permanent = 1)
dispatch.skipStep("xcustom", permanent = 1)

As a result, Anaconda does not bother detecting the video card and just runs. This was the source of the problem for the failure to load X. As such, the solution is to download the .iso, mount the .iso and copy the contents into another directory for modification:

mkdir /tmp/kernel-iso
mount -o loop kernel-4.2.1-0.x86_64.disk1.iso /tmp/kernel-iso
cp -av /tmp/kernel-iso /tmp/kernel-iso-new

modify these lines in the ks.cfg file as follows:

dispatch.skipStep("handleX11pkgs", skip = 0)
dispatch.skipStep("videocard", skip = 0)
dispatch.skipStep("monitor", skip = 0)
dispatch.skipStep("xcustom", skip = 0)

remake the .iso:

mkisofs -o /tmp/kernel-4.2.1-new.iso -b isolinux/isolinux.bin \
-c isolinux/boot.cat -no-emul-boot -boot-load-size 4 \
-boot-info-table -J -R -V 'kernel disk1' .

burn it and run again.

Now, I managed to get until partitioning my hard disk. After I had set on how to partition the hard disk, Anaconda ask for Core Roll CD (that contains Area 51, Ganglia, SGE and etc). When I clicked next, this error was displayed :

Traceback (most recent call last) :
File "/var/tmp/anaconda-10.1.1.46//usr/lib/anaconda/gui.py", line 1074, in handleRenderCallback self.currentWindow.renderCallback()
File "/tmp/updates/usr/lib/anaconda/progress-gui.py", line 249 in renderCallback self.intf.icw.nextClicked()
File "/var/tmp/anaconda-10.1.1.46//usr/lib/anaconda/gui.py", line 789 in nextClicked self.dispatch.gotoNext()
File "/var/tmp/anaconda-10.1.1.46//usr/lib/anaconda/dispatch.py", line 171 in gotoNext self.moveStep()
File "/var/tmp/anaconda-10.1.1.46//usr/lib/anaconda/dispatch.py", line 239 in moveStep rc=apply(func, self.bindArgs(args))
File "/tmp/ksclass.py", line 1409 in RocksPreInstall
File "/tmp/ksclass.py", line 1397 in RocksReadComps
File "/tmp/ksclass.py", line 1332 in RocksGetRools
File "/tmp/ksclass.py", lone 1269 in downloadRoll

OSError:[Errno 2] No such file or directory :
"/mnt/sysimage/export/home/install/rolls/area51/4.2.1/i386"

My team managed to install the frontend node today. It is because we need to install automatic partitioning. I don't like automatic partitioning but I think I can endure it. But now the screen is hang. Let see if it can managed to load all of it. I'll post another updates later.