Archive
Posts from 2007
Running commands across multiple servers with clusterssh
I periodically need to perform repetitive maintenance operations (e.g., patching systems) on groups of servers, which typically requires me to run a similar set of commands on multiple hosts. To make my life easier, I use the super useful clusterssh utility to interactively run commands across a group of servers. Cluster SSH is super easy to configure, and uses the concept of a "cluster" to define a group of similar nodes. To get up and running with clusterssh, you will first need to run cssh with the "-u" option to generate a configuration file (this step is optional, but creating a config file reduces start up time): Once the configuration file is created, you can open the file in a text editor, and change the settings to fit your administration preferences…
$ read more →Making rc scripts chkconfig aware
After adding a new rc script to /etc/init.d/ on a RHEL 4 box last week, I was greeted with the following error when I ran chkconfig to create the /etc/rc0.d /etc/rc1.d /etc/rc2.d /etc/rc3.d /etc/rc4.d /etc/rc5.d /etc/rc6.d symbolic links: After a big of poking around, it look like chkconfig looks for a line similar to the following in each run control script: The values after the "chkconfig:" statement contain the runlevels to enable the script at, the value to use after the "S" in the start scripts, and the value to use after the "K" in the kill scripts. So 345 would cause the start script to be executed at run levels 3, 4 and 5, the start script would be named S99llc2, and the kill script would be named K50llc2. Shibby!
$ read more →Monitoring hardware RAID controllers with Solaris
I manage several V40Zs running Solaris 10, and these servers utilize the built-in hardware RAID controller. Siince the physical spindles are masked off from the operating system, using a tool like smartmontools to check disk health is not an option. Luckily Solaris shops with the raidctl utility, which provides insight into the status of both the controller and the disks that sit behind that controller: Since raidctl will display a disk fault when a drive fails, I run a shell wrapper from cron every fifteen minutes to check the RAID controller status. If the script detects a problem, it will send an email and generate a syslog entry to let folks know a problem exists…
$ read more →Limiting the size of Solaris tmpfs file systems
I had an application go nuts a week or two ago, and it filled up /tmp on one of my Solaris 10 hosts. Since /tmp is an in memory file system, you can only imagine the chaos this caused. :( To ensure that this never happens again, I modified the tmpfs entry in /etc/vfstab to limit tmpfs to 1GB in size: swap - /tmp tmpfs - yes size=1024m That will teach that pesky application. :)
$ read more →VxFS clear blocks mount option
While reading through the VxFS administrators guide last week, I came across a cool mount option that can be used to zero out file system blocks prior to use: "In environments where performance is more important than absolute data integrity, the preceding situation is not of great concern. However, for environments where data integrity is critical, the VxFS file system provides a mount -o blkclear option that guarantees that uninitialized data does not appear in a file." This is pretty cool, and a useful feature for environments that are super concerned about data integrity
$ read more →