Editing Ploop/readme

This document explains how to use user-space ploop utility for typical
use-cases.

== Getting help ==

All user-space ploop management operations are available via "ploop"
utility. Run it w/o args to get help:

 # ploop

Run with a cmd as the only arg to get cmd-specific help, e.g.:

 # ploop init

== Initialize image file ==

In-kernel ploop operates on image files of "raw" or "ploop1" format.
An image file should be created and initialized (entirely in user-space)
before asking in-kernel ploop to start using it.

To create and initialize image file of "ploop1" format:

 # ploop init -s 1g -f ploop1 /ploop.image

the same for "raw" format:

 # ploop init -s 1g -f raw /ploop.image

where /ploop.image is full path to new image file and 1g is block-device size
equal to 1GB. These commands succeed only if the file /ploop.image didn't
exist at the time of running "ploop init".

== Load modules ==

Further ploop commands assume that all necessary modules are loaded:

 # modprobe ploop
 # modprobe pfmt_ploop1
 # modprobe pfmt_raw
 # modprobe pio_direct

It's not always required to load both pfmt_ploop1 (support of ploop1 format) and
pfmt_raw (support of raw format). If we're going to use ploop1 image file,
loading pfmt_raw can be omitted. If we're going to use raw image file and have no
plans to snapshot it, pfmt_ploop1 can be omitted.

== Create device file ==

 # mknod /dev/ploop0 b 182 0

This example above uses 0 as minor number. In practice, it can be any multiple
of 16 in 0..1048575 range: 0, 16, 32, 48, etc.

Other minors (not multiple of 16) are reserved for partitions. E.g. if primary
partition number 1 is created on /dev/ploop0, /dev/ploop0p1 will have minor
equal to 1.

== Bind ploop to image file ==

Assuming that steps 2-4 were done, the following commands are used to "mount"
ploop device over image file.

For ploop1 format:

 # ploop mount -f ploop1 -d /dev/ploop0 /ploop.image

For raw format:

 # ploop mount -f raw -d /dev/ploop0 /ploop.image

Since this point, /dev/ploop0 is operable. One can read/write any data from/to
it (e.g. with "dd"), manipulate partition table on it (with "fdisk"), format
it with mkfs.ext4 and mount it on some mount-point. In the other words, since
now /dev/ploop0 can be used as any other ordinary block device.

== Snapshot ==

Let /dev/ploop0 be a ploop device running over /ploop.image (i.e. step 4 was
done) and /ploop-delta.image - some full path to non-existent file. Then the
command:

 # ploop snapshot -d /dev/ploop0 /ploop-delta.image

will create empty "ploop1" image file and register it in kernel ploop
forming "snapshotted" configuration top_delta-->base_delta where base_delta
is /ploop.image and top_delta is /ploop-delta.image.

Since now, all i/o targeted at /dev/ploop0 will change only top_delta.
Actually, while performing snapshot operation in kernel, ploop re-open
base_delta "read-only". So, when "ploop snapshot" completed, it's quite safe,
for example, to backup base_delta.

It's allowable to snapshot snapshotted configuration. Following example above,
the command:

 # ploop snapshot -d /dev/ploop0 /ploop-delta1.image

will form top_delta-->delta-->base_delta configuration where:

 top_delta := /ploop-delta1.image
 delta := /ploop-delta.image
 base_delta := /ploop.image.

All deltas in snapshotted configuration are enumerated in kernel ploop in
the natural order starting from 0 for base_delta:

 base_delta' number is 0
 delta' number is 1
 top_delta' number is 2.

This knowledge is useful for online merge below.

== Merge ==

Merge operation implies copying all new data from an upper delta to a lower
delta. In simplest case of delta2-->delta1 configuration, merge will copy
new data from delta2 to delta1. In case of deltaN-->...-->delta1 configuration.
merge will copy new data from <deltaN, ..., delta2> to delta1.

There are two types of merge: offline and online. "offline" means that we have
a bunch of stand-alone image files w/o kernel ploop running over them.
"online" means that we have kernel ploop running over them. It's OK to perform
online merge concurrently with other i/o activity (i.e. no need to stop ploop
device or freeze upper-layer apps).

=== Offline merge ===

ploop has no heruistic about determining raw/ploop1 format of image file. So,
in case of offline merge, user should specify the format of base_delta
explicitly. If it's ploop1, merge command looks like:

 # ploop merge /ploop-delta.image /ploop.image

This will merge /ploop-delta.image into /ploop.image. More than one source
delta can be specified:

 # ploop merge /ploop-d2.image /ploop-d1.image /ploop-d.image /ploop.image

This will merge /ploop-d2.image, /ploop-d1.image and /ploop-d.image into
/ploop.image.

When merge completed, source deltas can be deleted because all data that
was present in them has been copied to destination delta.

For raw format, "-f raw" should be added as option. E.g.:

 # ploop merge -f raw /ploop-d1.image /ploop.image

=== Online merge ===

In this case user should only specify ploop device and a range of deltas in the
form of LEVEL1..LEVEL2 where LEVEL1 and LEVEL2 should be non-negative integers
corresponding to in-kernel delta enumeration (see the end of 5th section above)
and LEVEL1 must be lesser than LEVEL2. When merge completed, source deltas are
deregistered from kernel ploop and can be deleted by user.

Let's consider simple snapshotted configuration as example:

 # ploop mount -f ploop1 -d /dev/ploop0 /ploop.image
 # ploop snapshot -d /dev/ploop0 /ploop-delta.image
 # ploop snapshot -d /dev/ploop0 /ploop-delta1.image

In this configuration the command:

 # ploop merge -d /dev/ploop0 -l 0..2

will merge /ploop-delta1.image and /ploop-delta.image into /ploop.image. Here
/ploop-delta1.image and /ploop-delta.image are source deltas and can be
deleted.

Alternatively, the command:

 # ploop merge -d /dev/ploop0 -l 0..1

will merge /ploop-delta.image into /ploop.image. Here /ploop-delta.image is
source delta and can be deleted.

The last example is:

 # ploop merge -d /dev/ploop0 -l 1..2

It will merge /ploop-delta1.image into /ploop-delta.image. Here
/ploop-delta1.image is source delta and can be deleted.

== Migration support ==

Assuming that /dev/ploop0 is ploop device running over /ploop.image,
/ploop1.image is the path to non-existent file and external_stop is some
executable script or binary, the following command:

 # ploop copy -s /dev/ploop0 -d /ploop1.image -F external_stop

will copy /home/ploop.image to /home/ploop1.image iteratively. external_stop
should be an utility that completely freeze all i/o targeted at ploop device.
For instance, container freeze.

From user view, the command above should be equivalent to "external_stop;
cp /home/ploop.image /home/ploop1.image". The benefit of "ploop copy" is
minimizing duration of freezed state: it will try to copy the most part
of data before calling external_stop and only some (hopefully small)
amount of data after external_stop.

It's also possible to split copy operation in two nodes:

 node01# ploop copy -d /ploop1.image
 node02# ploop copy -s /dev/ploop0 -F external_stop

In this case "ploop copy" on node02 will write data to standard output
in some special binary format and "ploop copy" on node01 will read data
from stndard input, parse that special format and store data in
/home/ploop1.image ("-s" stands for "source", "-d" stands for "destination").
This should work if standard output of node02 and standard input of node01
are bound via pipe or socket connection.

== Grow ploop device ==

"ploop grow" command is to extend image file as neccessary (offline or online)
and propogate new block-device size to linux kernel (in online case).

=== Offline grow ===

To grow image file offline, user should specify its format explicitly. The
default is "ploop1" format:

 # ploop grow -s 32g /ploop.image

will re-arrange "ploop1" image file /ploop.image to become 32GB size long.

For "raw" format, the command:

 # ploop grow -s 32g -f raw /ploop.image

will do the same.

=== Online grow ===

Assuming that /dev/ploop0 is runnung ploop device, the command:

 # ploop grow -s 32g -d /dev/ploop0

will re-arrange underlying image file, update internal kernel ploop structures
and propogate changes to linux kernel making ploop device 32GB size long.

If user had ext4 formatted and mounted on /dev/ploop0, ext4 fs can be extended
online (when "ploop grow" completed):

 # resize2fs /dev/ploop0 32g

== Ballooning ==

ploop doesn't support pure shrinking block-device size due to lack of online
shrink support in ext4. As a workaround, "ballooning" technique is proposed.
Ballooning operation consists of inflating special balloon file in user-space
(the file will be invisible for ordinary users, e.g. inside container),
loading fiemap info of inflated balloon to kernel, relocating blocks of
image file from the tail to the space specified by fiemap info and truncating
tail of image file.

Desired outcome is image file of smaller size. However, it's quite possible
that inflated balloon file will span only blocks that were never touched
before. They will look as "not allocated" space from kernel ploop view. In this
case nothing will be relocated and nothing truncated.

So, if balloon operation succeeded, it's only guaranteed that user of ploop
device won't be able to consume more space than initial block device size
minus size of inflated balloon. On the other hand, if user of block device
used a lot of space on it, then freed significant part of used space,
balloon operation will result in significant truncate of image file.

To enable ballooning, ext4 residing on ploop device should be mounted
with special "balloon_ino" option:

 # mount -t ext4 -o balloon_ino=12 /dev/ploop0 /mnt_ploop

where 12 is inode number of balloon file as reported by "ls -i".
(it's assumed that initially, e.g. while constructing container,
someone mounted ext4 on ploop device w/o balloon_ino option, then created
empty balloon file there, found out its inode number and saved it for the
future use)

Currently, only online ballooning is supported. The following command performs
this operation:

 # ploop balloon change -s 1g -d /dev/ploop0 -m /mnt_ploop

where 1g is desired new size of balloon file, /dev/ploop0 is ploop block
device, /mnt_ploop is mount-point where ext4 residing on /dev/ploop0 is
mounted to.

If balloon file was empty, the command above simply iflates it to become
1GB size. If it was non-empty but smaller than 1GB, that command extends it
to given size (1GB). If it was non-empty but larger that 1GB, that
command truncates it down to given size. If it was exactly 1GB size, the
command does nothing.

Along with "change" sub-command, "ploop balloon" supports a few auxiliary ones:

 # ploop balloon show -m /mnt_ploop

will show current ploop balloon size.

 # ploop balloon status -d /dev/ploop0 -m /mnt_ploop

will report current in-kernel status of maintainance like "merge in progress",
"grow in progress", "ballooning started", etc. This is useful because on the
one hand balloon operation can't be performed while merge or grow is in
progress, and on the other hand previous "ploop balloon" could be killed by
someone before its completion.

 # ploop balloon clear -d /dev/ploop0 -m /mnt_ploop

will flush stale in-kernel "BALLOON" state of maintainance. This is useful if
previous "ploop balloon" died early leaving in-kernel ploop locked.

 # ploop balloon complete -d /dev/ploop0 -m /mnt_ploop

will complete previously interrupted balloon operation. An expectation is that
user monitors exit status of ploop commands he/she runs in some way. If
user issued "ploop balloon change" and it was killed in the middle, the user
knows that it didn't complete with zero exit status. Then user shoud inquire
current maintainance state with "ploop balloon status" command, and, if it
reported "FBLOAD" or "RELOC", the user should use "ploop balloon complete"
before proceeding with any other maintainance operations (shanphsot, merge,
grow, balloon).

 # ploop balloon check -d /dev/ploop0 -m /mnt_ploop

will check whether existent balloon file was properly processed. This is useful
if previous "ploop balloon" was interrupted, but "ploop balloon status"
reports "OFF" or "BALLOON" maintainance state. In this case it's possible
that balloon file was inflated but no further processing happened.

"ploop balloon check" reports total number of free blocks in existent balloon
file. If it's not zero, the user should use the following command to repair
balloon:

 # ploop balloon repair -d /dev/ploop0 -m /mnt_ploop

This command does essentially the same as "ploop balloon change" but w/o
inflating balloon.