Certification Summary

The processes running on a system need random access memory (RAM), and when the system runs low on memory, some disk space is used in a way that the process views it as memory. This method of using disk space for memory is called virtual memory, and the disk space being used as memory is called physical swap space. In other words, virtual memory combines the swap space with memory. To avoid allocating unnecessary swap space, an address for the swap space that actually exists in the physical memory is assigned to the process when it asks for swap space. This is called virtual swap space. The SWAPFS file system is responsible for providing these virtual swap space addresses and translating them to real, physical swap space addresses when the system runs low on memory and the process needs to use the physical swap space. The TMPFS file system, a memory-based file system that stores its files in the /tmp directory in memory, is used to improve the performance of the applications. This system uses the space resources from the swap space. The mkfile and swap commands are used to manage the swap space.

The file systems are shared over the network by using the NFS service, which is automatically started when the system is booted. The nfsd daemon handles the file system requests from the client. A file system on the server that you need to share each time the system is booted should have a share command in the /etc/dfs/dfstab file. Because NFS service is managed by NFS, you can use the svcs command to determine the status of the service, and the svcadm command to enable and disable the service. The shared file systems can be used by the client machine by mounting them. For that purpose Solaris offers an automount SMF service called autofs, which is started automatically at boot time and has the daemon named automountd. The mount points on a client machine are mapped to the file systems on the server machines by three kinds of files on the client machine. These are the master maps file /etc/auto_master and the direct and indirect maps files defined in the auto_master file. The main steps in troubleshooting NFS services consist of: using the ping command to check the connectivity between the client and the server machines, using the rpcinfo command to check whether the client can access the NFS services running on the service machine, and using the rpcinfo command to check whether the NFS services are actually running.

Processes running on the system as well as the system itself do crash as a result of hardware failures or software errors. The problems that may have caused the crash can be diagnosed by examining the files that the crashes create: core dump files created when a process terminates abnormally and crash dump files created when the system crashes. These files can be managed by using the coreadm and dumpadm commands.

The network file system, and the virtual file systems help hide the boundaries from users. For example, a user does not have to worry about which machine has the data. From the end user's perspective, it is right there on the user's machine even if it is actually on another machine. There is another important boundary to consider—the limitation of having a file system on one disk. That is, a file system resides on a disk slice and hence cannot span across multiple disks. The volume management offered by Solaris eliminates this boundary for the end user as well. We explore this topic in the next chapter.

Inside the Exam

Comprehend

• Virtual memory is the storage space reserved on the hard disk for the system to use as memory when it runs low on physical memory. The disk space being used for the virtual memory is called swap space.

• The names of the direct maps file and the indirect maps files are defined in the master maps file: /etc/auto_master, An entry in the direct maps file provides the full path of the mount point The full path of the mount point in an entry in an indirect maps file is composed of the mount point defined in the indirect maps file and the mount point defined in the •', corresponding entry in the master maps file,

• Core dump files are created when processes running on the system terminate abnormally, whereas crash dump files are created when the system itself crashes.

Look Out

• When paging starts, the SWAPFS file system has to map the virtual swap space addresses to physical swap space addresses.

• The reserved swap space does not have any assigned location on the disk; the location is assigned when the process wants to write data and the reserved space is actually allocated.

• You don't need to repartition the disk to add secondary swap space by using the mkfile and swap commands.

• To share a file system, you must have issued the share command for the file system either manually or by making it an entry in the /etc/dfs/dfstab file, and the SMF server must be running.

• The default access to the shared file system is read/write for all clients.

• In the framework of automount service, only mount points are created at system startup times, and no file system is mounted on those points during the boot.

• Unlike mount, automount does not read the /etc/vfstab file for a list of file systems to mount. Remember that this file is specific to each computer.

Memorize

• TMPFS manages the swap space files, and SWAPFS manages the virtual swap space addresses.

• Virtual swap space is the sum of the actual physical swap space plus a portion of the available physical memory space.

• The swap devices (files and slices) are listed in the /etc/vfstab file.

• The swap slices and swap files are activated by the /sbin/Swapadd script each time the system is booted.

• The parameters for the nfsd daemon are in the /etc/default/nfs file,

• The coredump files are managed by the coreadm command and the crash dump information is managed by the dumpadm command; the configuration files for these two utilities are coreadm.conf and dumpadm.conf.