1dlmfs 2================== 3A minimal DLM userspace interface implemented via a virtual file 4system. 5 6dlmfs is built with OCFS2 as it requires most of its infrastructure. 7 8Project web page: http://ocfs2.wiki.kernel.org 9Tools web page: https://github.com/markfasheh/ocfs2-tools 10OCFS2 mailing lists: http://oss.oracle.com/projects/ocfs2/mailman/ 11 12All code copyright 2005 Oracle except when otherwise noted. 13 14CREDITS 15======= 16 17Some code taken from ramfs which is Copyright (C) 2000 Linus Torvalds 18and Transmeta Corp. 19 20Mark Fasheh <mark.fasheh@oracle.com> 21 22Caveats 23======= 24- Right now it only works with the OCFS2 DLM, though support for other 25 DLM implementations should not be a major issue. 26 27Mount options 28============= 29None 30 31Usage 32===== 33 34If you're just interested in OCFS2, then please see ocfs2.txt. The 35rest of this document will be geared towards those who want to use 36dlmfs for easy to setup and easy to use clustered locking in 37userspace. 38 39Setup 40===== 41 42dlmfs requires that the OCFS2 cluster infrastructure be in 43place. Please download ocfs2-tools from the above url and configure a 44cluster. 45 46You'll want to start heartbeating on a volume which all the nodes in 47your lockspace can access. The easiest way to do this is via 48ocfs2_hb_ctl (distributed with ocfs2-tools). Right now it requires 49that an OCFS2 file system be in place so that it can automatically 50find its heartbeat area, though it will eventually support heartbeat 51against raw disks. 52 53Please see the ocfs2_hb_ctl and mkfs.ocfs2 manual pages distributed 54with ocfs2-tools. 55 56Once you're heartbeating, DLM lock 'domains' can be easily created / 57destroyed and locks within them accessed. 58 59Locking 60======= 61 62Users may access dlmfs via standard file system calls, or they can use 63'libo2dlm' (distributed with ocfs2-tools) which abstracts the file 64system calls and presents a more traditional locking api. 65 66dlmfs handles lock caching automatically for the user, so a lock 67request for an already acquired lock will not generate another DLM 68call. Userspace programs are assumed to handle their own local 69locking. 70 71Two levels of locks are supported - Shared Read, and Exclusive. 72Also supported is a Trylock operation. 73 74For information on the libo2dlm interface, please see o2dlm.h, 75distributed with ocfs2-tools. 76 77Lock value blocks can be read and written to a resource via read(2) 78and write(2) against the fd obtained via your open(2) call. The 79maximum currently supported LVB length is 64 bytes (though that is an 80OCFS2 DLM limitation). Through this mechanism, users of dlmfs can share 81small amounts of data amongst their nodes. 82 83mkdir(2) signals dlmfs to join a domain (which will have the same name 84as the resulting directory) 85 86rmdir(2) signals dlmfs to leave the domain 87 88Locks for a given domain are represented by regular inodes inside the 89domain directory. Locking against them is done via the open(2) system 90call. 91 92The open(2) call will not return until your lock has been granted or 93an error has occurred, unless it has been instructed to do a trylock 94operation. If the lock succeeds, you'll get an fd. 95 96open(2) with O_CREAT to ensure the resource inode is created - dlmfs does 97not automatically create inodes for existing lock resources. 98 99Open Flag Lock Request Type 100--------- ----------------- 101O_RDONLY Shared Read 102O_RDWR Exclusive 103 104Open Flag Resulting Locking Behavior 105--------- -------------------------- 106O_NONBLOCK Trylock operation 107 108You must provide exactly one of O_RDONLY or O_RDWR. 109 110If O_NONBLOCK is also provided and the trylock operation was valid but 111could not lock the resource then open(2) will return ETXTBUSY. 112 113close(2) drops the lock associated with your fd. 114 115Modes passed to mkdir(2) or open(2) are adhered to locally. Chown is 116supported locally as well. This means you can use them to restrict 117access to the resources via dlmfs on your local node only. 118 119The resource LVB may be read from the fd in either Shared Read or 120Exclusive modes via the read(2) system call. It can be written via 121write(2) only when open in Exclusive mode. 122 123Once written, an LVB will be visible to other nodes who obtain Read 124Only or higher level locks on the resource. 125 126See Also 127======== 128http://opendlm.sourceforge.net/cvsmirror/opendlm/docs/dlmbook_final.pdf 129 130For more information on the VMS distributed locking API. 131