# This file contains configuration information for venus, the Coda # client. Most of the things specified here can be overridden by # command-line flags. # # Which are the root servers we should contact for volume location # information. When specifying multiple servers, uses ',' to separate, # no extra spaces! (f.i. server1,server2,server3) # rootservers=testserver.coda.cs.cmu.edu # # Which are the root servers we should contact for user authentication. # When not specified, the rootservers are contacted. # #authservers=testserver.coda.cs.cmu.edu # # What should the size of the local cache be in 1k blocks. If this is # not specified or `0' the default value of 40000 (40MB) is chosen. # Minimum value is 2048. # cacheblocks=100000 # # How many files should the venus cache hold. If this is not specified # or `0' the default value is calculated as 'cacheblocks / 24'. # Minimum value is 85. # #cachefiles=0 # # How many modification log entries should venus have. If this is not # specified or `0' the value is calculated as 'cacheblocks / 6'. # (with the default number of cachefiles this is '4 * cachefiles') # Minimum value is 256. # #cml_entries=0 # # How many hoard database entries should venus have. If this is not # specified or `0' the value is calculated as 'cacheblocks / 48'. # (with the default number of cachefiles this is 'cachefiles / 2') # Minimum value is 42. # #hoard_entries=0 # # Which file should receive venus's stderr output. # (default is /usr/coda/etc/console). # #errorlog=/usr/coda/etc/console # # What is the uid of the primary user of this system. The primary user # has the rights to modify the entries in the hoard database. Normally # venus can successfully figure this out by looking which user has the # `console', or is logged in on tty1, the first virtual console. Root is # always considered `primary user'. # If a non-zero uid is specified here, venus will not attempt to guess # who the primary user is. # #primaryuser=0 # # What are the locations of the rvm log and data files or partitions. # (default /usr/coda/{LOG,DATA} # # CAUTION: these are _unrelated_ to the server's rvm log and data. Also # when specifying partitions, do NOT specify any partions which might # have an existing filesystem, as you WILL lose all data during # initialization. # #rvm_log=/usr/coda/LOG #rvm_data=/usr/coda/DATA # # Where should venus keep it's pool of cached files. # (default is /usr/coda/venus.cache) # #cachedir=/usr/coda/venus.cache # # Where should venus store snapshots of the modification logs. This is # primarily to provide some safe fallback when venus dies or messes up # the CML during reintegration. (default is /usr/coda/spool) # #checkpointdir=/usr/coda/spool # # Where does venus store it's pidfile # #pid_file=/usr/coda/venus.cache/pid # # What file does vutil use to control venus # #run_control_file=/usr/coda/venus.cache/VENUS_CTRL # # Venus log file # #logfile=/usr/coda/etc/venus.log # # Mariner socket, # On platforms that support unix domain sockets this is the socket that # codacon and spy use to talk to venus. On other platforms they will # connect to venus through tcp port 2430. # #marinersocket=/usr/coda/spool/mariner # # Dont use RVM, # Avoid using persistent storage for metadata. Only useful on diskless # clients or small handhelds when the venus.cache is stored in a ramdisk # or ram filesystem. Beware, if you enable this option, venus will loose # all not reintegrated modifications when it dies. # #dontuservm=0 # # Map private, # Whether to use a private mmaping of the RVM data file instead of a # anonymous mmap. Using a private mmap will reduce startup times, as # missing pages are paged in when they are accesses. Also swap usage is # reduced as any unmodified pages do not have to be backed up by swap # memory when memory gets tight. Private mappings do not work when rvm # data is stored on a raw partition. Set to 1 to use private mappings, # 0 to use anonymous mappings. # mapprivate=1 # # Mountpoint, # Where the coda filesystem is mounted, many application still depend on # this being /coda. I am unsure whether it matters much on Windows95/98, # but we normally use N: for those. # #mountpoint=N: #mountpoint=/coda # # Kernel device, # The character device used by venus to communicate with the kernel module. # Standard name is /dev/cfs0, ignored by the Win95/98 port. Maybe we will # switch to using /dev/coda0 at some point to avoid name conflicts. # #kerneldevice=/dev/cfs0 # # What is the name of the rootvolume. Normally this is queried from the # rootservers, only in very unusual cases does this need to be set. # #rootvolume= # # The masquerade option enables SFTP to use the same ports as RPC2 for # bulk data transfers. This allows a Coda client to connect to servers # from behind a masquerading firewall. The rpc2 library on both the # client and the servers must be at least version 1.9 for this to work. # masquerade=1 # # On NT with Cygwin, venus names are different than the NT real names. # Venus names are relative to the root of the cygwin tree. Replace # the C:/cygwin with the actual root directory of your cygwin installation. # Do not remove the /??/ part of the prefix. # #cache_prefix=/??/C:/cygwin # # Maximum number of fids & versionvectors that will be piggybacked with a # ValidateAttrs RPC2 call. The default (and maximum) is 50, however the # resulting outgoing packet is around 2900 bytes which will be fragmented by # the IP layer. In some cases this IP fragmentation is not wanted (unreliable # networks, IPSec, and in the future IPv6 networks). Setting this number to 15 # piggybacked validations limits the outgoing ValidateAttr packet size to # about 1032 bytes, and 21 validations would result in a 1452 byte packet. # #validateattrs=50 # # How many seconds between checks whether the servers are still alive. The # default is 12 minutes. However masquerading firewalls will typically forget # about UDP connections sooner. For instance, connection tracking in Linux's # netfilter only remembers UDP streams for about 3 minutes, so 150 seconds # would be a good choice here. # # If you are going through a firewall and regularily see messages that the # servers nak'ed lowering this timeout could very well help a lot. # #serverprobe=720