This chapter explains and describes the Demand-Attach File Server (DAFS for short), a new feature in OpenAFS 1.6. A few things are different when setting up and maintaining a DAFS fileserver vs a traditional fileserver, so it is important to know which one you are running. If you are trying to decide whether or not to run DAFS on a fileserver, note that you can switch between DAFS and the traditional fileserver fairly easily at any time. Aside from performance differences, the two fileservers generally look and act similarly, so choosing one over the other in the beginning should not hurt you later on. So, if you are not overly concerned with performance, you can just pick one and follow the directions for it, and you should be fine. DAFS changes many things with how the fileserver and other server processes access volumes on disk. Most of these changes are only of concern to developers, as there are no visible differences to users or administrators. A few changes are only of concern to administrators while debugging a problem, and only one or two changes affect the day-to-day administration of a fileserver. See the sections on DAFS Binaries and Salvaging for the main functional differences. Among other things, DAFS provides a new way for the fileserver and other server programs to access volumes. It does not change how file data is accessed inside a volume (like namei or inode do), but rather how a volume as a whole is accessed by the fileserver. When a traditional fileserver is started, it will locate and attach all AFS volumes it can find on all AFS partitions on the server. This attachment process involves reading the volume header and setting a flag to say that the volume is in use. On a clean shutdown, the fileserver detach all volumes it attached by clearing that flag. If the fileserver encounters a volume where the flag is already set before it attached the volume, it will know that volume was not detached cleanly, and that it needs to be salvaged. This process of attaching and detaching at startup and shutdown takes time. When fileservers start to have thousands of volumes, it can take several minutes to start or cleanly shutdown. Additionally, when a fileserver is shutdown uncleanly, all of the volumes on the server must be salvaged before the fileserver can be brought up again. Even volumes that nobody has ever accessed must be salvaged. To improve this situation, DAFS does not attach all volumes on startup. Instead, volumes are only attached when they are first accessed, and are detached if idle for a long period of time (see the -vlru* options for the fileserver; this is configurable but is generally several hours). Thus, after an unclean shutdown with DAFS, only the volumes in use at the time of the shutdown need to be salvaged. In addition, the fileserver does not need to wait for salvaging to finish before starting up. Instead of salvaging everything before the fileserver starts, the fileserver causes salvages to be issued on a volume when a damanged volume is accessed (called "demand-salvages"). The result of all of this is that a fileserver running DAFS can generally be started and stopped in a manner of seconds, where a traditional fileserver could take dozens of minutes, or even hours in the case of an unclean shutdown. OpenAFS ships with binaries for DAFS and non-DAFS alongside each other. Programs that exist in both DAFS and non-DAFS variants typically have a da prefix to indicate the DAFS variant. For example, dafileserver is the DAFS version of the traditional fileserver binary. Similarly for davolserver and volserver, dasalvager and salvager, and even some other tools like dafssync-debug and fssync-debug. A DAFS-enabled fileserver will run the dafs bnode in bosserver, instead of the traditional fileserver's fs bnode. The dafs bnode runs four different binaries as opposed to the fs bnode's three. Three of the programs in the dafs bnode are just DAFS equivalents of the corresponding fs bnode programs, but the fourth one is an entirely new daemon that has no non-DAFS equivalent: the salvageserver. Traditional binary DAFS binary /usr/afs/bin/fileserver /usr/afs/bin/dafileserver /usr/afs/bin/volserver /usr/afs/bin/davolserver No equivalent /usr/afs/bin/salvageserver /usr/afs/bin/salvager /usr/afs/bin/dasalvager With a traditional fileserver, salvages usually occur in two situations: The fileserver shuts down uncleanly, and when brought back up, all partitions are salvaged before the fileserver is available. A volume experiences some corruption after it has been brought online, and an administrator manually schedules a salvage for an individual volume with bos salvage. Usually the way you notice this is that the fileserver noticed a volume has become corrupt and has taken it offline. With DAFS, neither of these occur in normal operation. With DAFS, a volume will be salvaged automatically when a problem is detected in a volume. This occurs whether the fileserver detects the volume was not detached cleanly, or if the fileserver detects corrupt volume data. In normal DAFS operation, you should not need to ever run bos salvage. However, if you suspect a bug, or that there is corruption in a volume that the fileserver has not detected, you can run bos salvage to manually issue a salvage. All salvages scheduled from the salvageserver will be logged in /usr/afs/logs/SalsrvLog, in the same format as salvages for traditional fileservers are logged. If you issue a whole-partition or whole-server salvage, the logs for that will still be located in /usr/afs/logs/SalvageLog. If you have an existing traditional fileserver, you can convert it to DAFS fairly easily. The conversion is in-place, but there is at least a small amount of downtime involved even if nothing goes wrong, since the fileserver processes must be restarted. If you want to avoid any downtime, move any volumes on the fileserver to another fileserver before performing the conversion. If the fileserver is currently running a version of OpenAFS earlier than 1.6.0, upgrade the fileserver to a version in the 1.6 series or later. This is not strictly necessary, but if you encounter problems here, it is helpful to catch them before converting to DAFS so as to isolate the problem. If you do not upgrade the fileserver before converting to DAFS, install the new OpenAFS binaries somewhere on the server. If the current bosserver process running on the fileserver is older than OpenAFS 1.6.0, you must upgrade it first. This can be done by replacing the bosserver binary and running bos restart -bosserver. This will cause all AFS processes controlled by the bosserver to be restarted. You can do this at the same as upgrading the fileserver in the previous step if desired, to avoid restarting the fileserver twice. Shutdown and stop the old fs bnode: # bos stop <machine name> fs -wait -cell <cell name> Create and start the new dafs bnode. # bos create <machine name> dafs dafs \ -cmd /usr/afs/bin/dafileserver <dafileserver options> \ -cmd /usr/afs/bin/davolserver <davolserver options> \ -cmd /usr/afs/bin/salvageserver <salvageserver options> \ -cmd /usr/afs/bin/dasalvager <dasalvager options> \ -cell <cell name> You can use the same options for the dafileserver process as you did for the fileserver process, and the same options for davolserver as you did for volserver. You can also use most of the same options for the salvageserver and dasalvager processes as you did for the salvager process; see their respective man pages. However, if you are upgrading from the 1.4.x series of OpenAFS, be aware that the -DontSalvage option does not exist anymore for the salvager (with or without DAFS). Also note that the dafileserver process has many more options to tune dafs-specific parameters. See the dafileserver man page for information about them. After you have performed these steps, switching back and forth between running a DAFS and a traditional fileserver is as simple a running bos stop on one bnode, and bos start on the other. Once you are confident that the DAFS processes are running smoothly and you do not anticipate switching back to the traditional fileserver, you can bos delete the fs bnode.