.TH "md_src_plugins_fstab_README" 3elektra "Sun May 29 2016" "Version 0.8.14" "Elektra" \" -*- nroff -*-
.ad l
.nh
.SH NAME
md_src_plugins_fstab_README \- README 

.IP "\(bu" 2
infos = Information about FSTAB plugin is in keys below
.IP "\(bu" 2
infos/author = Markus Raab elektra@libelektra.org
.IP "\(bu" 2
infos/licence = BSD
.IP "\(bu" 2
infos/needs =
.IP "\(bu" 2
infos/provides = storage
.IP "\(bu" 2
infos/placements = getstorage setstorage
.IP "\(bu" 2
infos/recommends = struct type path
.IP "\(bu" 2
infos/description = Parses files in a syntax like /etc/fstab file
.PP
.PP
This plugin is an implementation of a parser and generator of the /etc/fstab file\&.
.PP
.SS "old fstab Entries"
.PP
(Deprecated, remove this section after it is reimplemented in the new way)
.PP
For each device in fstab elektra will store the following keys: pseudoname/device pseudoname/mpoint pseudoname/type pseudoname/options pseudoname/dumpfreq pseudoname/passno
.PP
Each represents a column in fstab\&.
.PP
The pseudoname can be any name for setting keys, the will be generated when getting keys, so don't expect the same name\&.
.PP
the directory / will be called rootfs
.PP
all swap devices will be called swapXX with a number from 00 on for XX
.PP
otherwise the mountpoint without the '/' character will be used\&.
.PP
At the other point there is the issue with the pseudonames, you can't rely on the pseudoname you have set\&.
.PP
The biggest issue is that you can't change or delete existing entries\&. All entries you set will be appended to the other filesystems\&.
.PP
So if you get the filesystems and change the type of the filesystem of the rootfs and set it again the resulting fstab will be like: /dev/sda6 / ext3>---- >----defaults,errors=remount-ro 0 1 /dev/sda6 / jfs>---- >----defaults,errors=remount-ro 0 1
.PP
which will be not like you desired!
.PP
setmntent is used, so it is only conforming to BSD 4\&.3 and linux and you can't use any comments\&.
.PP
.SS "new fstab Entries"
.PP
Specification: 
.PP
.nf
[/_]
type = array
explanation = the name of the key is the mountpoint (so the former
  mpoint is not needed); the value is the number of entries in the
  array
[/_/#]
explanation = an entry in the array
[/_/#/device]
[/_/#/type]
[/_/#/options]
[/_/#/dumpfreq]
[/_/#/passno]

.fi
.PP
.PP
Example: A fstab that looks like: 
.PP
.nf
/dev/sr0        /media/cdrom   udf,iso9660 user,noauto     0       0

.fi
.PP
.PP
would have a key name that is an array (so the value is the number of children, in this case 1): 
.PP
.nf
system/filesystems/\/media\/cdrom

.fi
.PP
.PP
with the array entry: 
.PP
.nf
system/filesystems/\/media\/cdrom0/#0/

.fi
.PP
.PP
So when following line is added 
.PP
.nf
/dev/sr0        /media/cdrom   ramfs user,noauto     0       0

.fi
.PP
.PP
Implementation hint: use \fBkeyAddBaseName()\fP to get escaping of /, then add array items below it
.PP
If a mount point exists more than once (that could be proc, swap or overlay mountpoints) the array below gets incremented (otherwise #0 is used for every unique entry)\&.
.PP
The order of the array must, of course, be preserved\&. Other lines may be reordered for now, a proper 'order' could be done later\&.
.PP
Spaces in the names are replaced by \\040 in the fstab\&.
.PP
.SS "Example"
.PP
Mount the plugin: 
.PP
.nf
    kdb mount /etc/fstab system/filesystems fstab struct type path

.fi
.PP
.PP
.SS "Note"
.PP
The status of this plugin is experimental\&.