[WIP] New repository system

Instead of the old domains this is the new rpository system. To add
another repo type add a module to storageRepository/engines/ and add

ENGINE_INFO = ("typeName", createFunc, instanceFunc)

This makes adding new domain types by 3rd parties easier.

Change-Id: I21fc36b33720da9529df40ddf6297fcb4fd4f3bc

Extracted image related functinality to another class

This will allow me to later have different image manipulator
implementation in the pool.

Change-Id: Ic29dc8d964f0c2e56329e7044a3b2f48a9f2e193

Move domain monitor out of sp.py

Domain monitoring logic should be seperated from the core pool logic to
reduce StoragePool class complexity

Change-Id: I18f30541a3466c532b2fd6144cd54b46545aff0c

rename vars

Change-Id: I93c2a70323ca1da7d03fdb530325936c359e324e

Fix __cleanupStoragePool()

Change-Id: I5ff24dde33e5a73fab4ab3c848d18071643af3a8

dir is a builtin, overriding it is not nice

Change-Id: I292cd9d5b47bb1051b765c02b55f351807316ea7

Fixed device mapper failing on partition check for non dm devices

Change-Id: I9b4713d7e0ee04722e82887378f448243241c631

BZ#751629 - Avoid migrate master if MSD is obsolete.

In addition deactivate the (block) domain if it's unreloadable.

Change-Id: I995b86f62ee9026875b966d66f216d295aac9eba
Reviewed-on: http://gerrit.usersys/1091
Reviewed-by: Ayal Baron
Tested-by: Haim Ateya <hateya@redhat.com>
Reviewed-by: Igor Lvovsky <ilvovsky@redhat.com>

vds_bootstrap: fix usage string

Change-Id: I8950764778dc61d04caeed4c74e792d52bba3378
Reviewed-on: http://gerrit.usersys.redhat.com/982
Reviewed-by: Dan Kenigsberg <danken@redhat.com>
Tested-by: Dan Kenigsberg <danken@redhat.com>

fix typo in trying to copy iscsid.conf template

Change-Id: I211a97dbd4c945e3ff30d4b42d072fd7d92761c2
Reviewed-on: http://gerrit.usersys.redhat.com/1088
Tested-by: Dan Kenigsberg <danken@redhat.com>
Reviewed-by: Dan Kenigsberg <danken@redhat.com>

Fixed SPM_mailbox attribute errror

Change-Id: Id3ccdf040248fee90c70803c6f0bf758e41d8568
Reviewed-on: http://gerrit.usersys.redhat.com/1079
Reviewed-by: Dan Kenigsberg <danken@redhat.com>
Tested-by: Dan Kenigsberg <danken@redhat.com>

BZ#749151 revive Vm ticket just before migration

We set a new expiry time on the spice password of the source qemu. That is
copied to the destination qemu as migration begins. We thus give spice client
two minutes to connect to destination to facilitate seamless migration.

v2:
- revive password only when spice client is connected
- revive password, do not remove it completely

v3:
- do not disconnect current client while reviving password.

Change-Id: Ibc7c0347354cf8f943c787dd621ce7b0eb78ef55
Reviewed-on: http://gerrit.usersys.redhat.com/1086
Reviewed-by: Igor Lvovsky <ilvovsky@redhat.com>
Reviewed-by: Dan Kenigsberg <danken@redhat.com>
Tested-by: Dan Kenigsberg <danken@redhat.com>

Revert "BZ#727602 setVmTicket: never expire ticket"

This reverts commit fd1f8a7a117067f1733898734b1c97bdf75d18bd.

Conflicts:

vdsm/libvirtvm.py

Change-Id: I7624e614576903e39ca0947a4f536433db944332
Reviewed-on: http://gerrit.usersys.redhat.com/1085
Reviewed-by: Igor Lvovsky <ilvovsky@redhat.com>
Tested-by: Dan Kenigsberg <danken@redhat.com>
Reviewed-by: Dan Kenigsberg <danken@redhat.com>

Move iscsid.conf template outside of the source file

Change-Id: I59ab2a951ddf15941218524ad7eb5bf94295023e
Reviewed-on: http://gerrit.usersys.redhat.com/1034
Reviewed-by: Dan Kenigsberg <danken@redhat.com>
Tested-by: Dan Kenigsberg <danken@redhat.com>

BZ#733000 vdsmd: upstart hates soflinks

Copy libvirt's upstart job definition instead of linking it.

Change-Id: I5c4f2a9c312c0f4fca984ff9bf07fac754bc6481
Reviewed-on: http://gerrit.usersys.redhat.com/1062
Reviewed-by: Dan Kenigsberg <danken@redhat.com>
Tested-by: Dan Kenigsberg <danken@redhat.com>

isMounted() now handles / at end of export in certain distros

Change-Id: I1cbba23b8a2d89a037ba809a86591ab33239ce24
Reviewed-on: http://gerrit.usersys.redhat.com/1072
Reviewed-by: Igor Lvovsky <ilvovsky@redhat.com>
Tested-by: Haim Ateya <hateya@redhat.com>
Reviewed-by: Saggi Mizrahi <smizrahi@redhat.com>

Changed default bridge name to engine

Change-Id: Iad298199b273753f921a2abfef16567d3e5a6033
Reviewed-on: http://gerrit.usersys.redhat.com/1071
Reviewed-by: Igor Lvovsky <ilvovsky@redhat.com>
Reviewed-by: Saggi Mizrahi <smizrahi@redhat.com>
Tested-by: Saggi Mizrahi <smizrahi@redhat.com>

Fixed reconstruct master

Change-Id: Ic7aa064dc07f2b26b3681be52a07901da197200b
Reviewed-on: http://gerrit.usersys.redhat.com/1070
Reviewed-by: Igor Lvovsky <ilvovsky@redhat.com>
Reviewed-by: Saggi Mizrahi <smizrahi@redhat.com>
Tested-by: Saggi Mizrahi <smizrahi@redhat.com>

Fix missing arg in sp.StoragePool() call.

Fix reconstruct master flow.

Change-Id: I00668572ca1aa4484171049393d4788046707c71
Reviewed-on: http://gerrit.usersys.redhat.com/1068
Tested-by: Dan Kenigsberg <danken@redhat.com>
Reviewed-by: Dan Kenigsberg <danken@redhat.com>

Fix unsecured on forceFreeSpm().

Change-Id: I4d224cd739e3bd13c5bcd9a8ed17818b8211187f
Reviewed-on: http://gerrit.usersys.redhat.com/1067
Tested-by: Dan Kenigsberg <danken@redhat.com>
Reviewed-by: Dan Kenigsberg <danken@redhat.com>

Fix lock release and reversed SPM logic.

Change-Id: I3e2335163e45dc601d6779b45beb03d67e997692
Reviewed-on: http://gerrit.usersys.redhat.com/1066
Reviewed-by: Dan Kenigsberg <danken@redhat.com>
Tested-by: Dan Kenigsberg <danken@redhat.com>

configure.ac: find scsi_id on Fedora 16, too.

Change-Id: Ic2b276931ddb31d01f7438928f931e3b951d3b56
Reviewed-on: http://gerrit.usersys.redhat.com/1065
Tested-by: Haim Ateya <hateya@redhat.com>
Reviewed-by: Dan Kenigsberg <danken@redhat.com>

Add the systemd support for bootstrap

Change-Id: I118256a15ddf442dc96168ab04db5b89627c83e3
Reviewed-on: http://gerrit.usersys.redhat.com/1043
Reviewed-by: Dan Kenigsberg <danken@redhat.com>
Tested-by: Dan Kenigsberg <danken@redhat.com>

RPM changes for Fedora

Change-Id: I5cc880f8579b254f49981d88f4045ac124559458
Reviewed-on: http://gerrit.usersys.redhat.com/1023
Reviewed-by: Dan Kenigsberg <danken@redhat.com>
Tested-by: Dan Kenigsberg <danken@redhat.com>

Recreate the run paths when needed

On some systems the run directory might be mounted as tmpfs (eg: on
Fedora). The daemon will refuse to start at next reboot if the paths
aren't recreated.

Change-Id: I6eee71ce11e47a64bf825089445d6ef4ec504d82
Reviewed-on: http://gerrit.usersys.redhat.com/1030
Reviewed-by: Dan Kenigsberg <danken@redhat.com>
Tested-by: Dan Kenigsberg <danken@redhat.com>

Convert the OrderedDict from namedtuple._asdict()

Since python 2.7 the namedtuple._asdict() method returns an
OrderedDict, therefore we need to convert it to a regular
dictionary before marshalling it on xmlrpc.

Change-Id: I5582f77fbfc6961dcada3f733cb9353dee53cf53
Reviewed-on: http://gerrit.usersys.redhat.com/1029
Reviewed-by: Igor Lvovsky <ilvovsky@redhat.com>
Reviewed-by: Dan Kenigsberg <danken@redhat.com>
Tested-by: Dan Kenigsberg <danken@redhat.com>

Related to 726960 - Restore higher version check.

This check was removed in commit 85b58b7d9 since the manager
can't send a lower pool version number.
In spite of this some tests stress this condition.
Restoring it in order to fix tests results.

Change-Id: I9c7258361589fac7dc94361b88cabdf73cf48a83
Reviewed-on: http://gerrit.usersys.redhat.com/927
Reviewed-by: Dan Kenigsberg <danken@redhat.com>
Tested-by: Dan Kenigsberg <danken@redhat.com>

use correct path for fencing agent

There's probably a nicer way to do it with autoconf, avoiding
os.path.dirname. Comments appreciated.

Change-Id: I346c8667ba3f822546f28c4f0edbdb6443d293f5
Reviewed-on: http://gerrit.usersys.redhat.com/1059
Reviewed-by: Saggi Mizrahi <smizrahi@redhat.com>
Reviewed-by: Dan Kenigsberg <danken@redhat.com>
Tested-by: Dan Kenigsberg <danken@redhat.com>

Fix call to getSpmRole()

Change-Id: I77aaa4aac2924c3df5f6d1c9a974dc94e7b0fcaf
Reviewed-on: http://gerrit.usersys.redhat.com/1061
Reviewed-by: Dan Kenigsberg <danken@redhat.com>
Tested-by: Dan Kenigsberg <danken@redhat.com>

BZ#748466 bootstrap: report vendor properly

Do not compare a string to a function, you are bound to fail.

Change-Id: Id9be847f4a59e2bea4db87cdbfa0fd6abe48be9d
Reviewed-on: http://gerrit.usersys.redhat.com/1060
Tested-by: Dan Kenigsberg <danken@redhat.com>
Reviewed-by: Igor Lvovsky <ilvovsky@redhat.com>

BZ#746766 BZ#747337: bootstrap: download selinux module before usage

Change-Id: I53e964ee4537831fa07576f0bda7cecfc96eae27
Reviewed-on: http://gerrit.usersys.redhat.com/1047
Reviewed-by: Igor Lvovsky <ilvovsky@redhat.com>
Tested-by: David Naori <dnaori@redhat.com>
Reviewed-by: Dan Kenigsberg <danken@redhat.com>

BZ#744755 Don't wait for tasks on shutdown

During shutdown we can't wait for all tasks to complete because some
of them might be stuck in D state if the storage is unreachable (eg:
createVolume preallocated on NFS).

Change-Id: I832517a7a336751ff6b3096d12d59e643bcbbe02
Reviewed-on: http://gerrit.usersys.redhat.com/1048
Reviewed-by: Igor Lvovsky <ilvovsky@redhat.com>
Tested-by: Dan Kenigsberg <danken@redhat.com>
Reviewed-by: Dan Kenigsberg <danken@redhat.com>

BZ#748222 netinfo: report operstate, not adminstate

Until commit 4153e08acb8ae23ef79 (fixing BZ#726359), when reporting if a
network inteface is up or down, Vdsm was considering the operative status.
That was the correct behavior, since users do not really care that they have
configered the nic to be up while the line is cut off by evil mice (or quality
engineers).

Change-Id: I32b74a57d96ba0feea4816749c6c45ea34695d2a
Reviewed-on: http://gerrit.usersys.redhat.com/1057
Reviewed-by: Igor Lvovsky <ilvovsky@redhat.com>
Reviewed-by: Dan Kenigsberg <danken@redhat.com>
Tested-by: Dan Kenigsberg <danken@redhat.com>

BZ#725540 vdsClient: check return code first

client should not assume that its data exists in the response before verifying
that the request was served without error.

Change-Id: Iffa0c0eb6b50b321f97a578994d7cf48bdae2974
Reviewed-on: http://gerrit.usersys.redhat.com/1031
Reviewed-by: Igor Lvovsky <ilvovsky@redhat.com>
Tested-by: Dan Kenigsberg <danken@redhat.com>

Fix bad rebase

Change-Id: Icaad8cd5214d78d0265b743c5f70d57d17be990a
Reviewed-on: http://gerrit.usersys.redhat.com/1058
Reviewed-by: Dan Kenigsberg <danken@redhat.com>
Tested-by: Dan Kenigsberg <danken@redhat.com>

BZ#729933 - Redefine public_activateStorageDomain()

Change-Id: Ic85953653c98c717a3c877fca80c2ca13cdd6547
Reviewed-on: http://gerrit.usersys.redhat.com/1014
Tested-by: Dan Kenigsberg <danken@redhat.com>
Reviewed-by: Dan Kenigsberg <danken@redhat.com>

Switched secure semantics to exclude rather then include

Change-Id: I95cd3e8bebcafe67dceb174d3e24623b7b5aa878
Reviewed-on: http://gerrit.usersys.redhat.com/1022
Reviewed-by: Dan Kenigsberg <danken@redhat.com>
Tested-by: Dan Kenigsberg <danken@redhat.com>

recoveryMode param is never used

Change-Id: Ifb7b5b491ad9899bd433ebaadc55cf6d46575edc
Reviewed-on: http://gerrit.usersys.redhat.com/1015
Reviewed-by: Dan Kenigsberg <danken@redhat.com>
Tested-by: Dan Kenigsberg <danken@redhat.com>

Merged SPM to StoragePool and HSM

SPM was an unholy union of two paradigms and was making everything hard.
This is an attempt to put everything in order

Change-Id: Ia016ba62a9436dd4f655bddc0921d60f5211781b
Reviewed-on: http://gerrit.usersys.redhat.com/1006
Reviewed-by: Dan Kenigsberg <danken@redhat.com>
Tested-by: Dan Kenigsberg <danken@redhat.com>

Made SPM_Mailmonitor independent from SPM

Change-Id: I4b0df9d46c3c71d04f0a6d69475651ddba0f84b1
Reviewed-on: http://gerrit.usersys.redhat.com/1000
Reviewed-by: Dan Kenigsberg <danken@redhat.com>
Tested-by: Dan Kenigsberg <danken@redhat.com>

Made spm securing logic more generic

 - Security is no longer system wide but instance wide
 - Security can be used on any class
 - Security override is supported by the subsystem and doesn't require
   trickery

Change-Id: I2e16feafc03d6efe5c904b18d4d27177dd14b905
Reviewed-on: http://gerrit.usersys.redhat.com/991
Reviewed-by: Dan Kenigsberg <danken@redhat.com>
Tested-by: Dan Kenigsberg <danken@redhat.com>

The start of exceptions being independent from the interface

Change-Id: Ic85a93ff0aa42369db477bfef2d7a340851cae73
Reviewed-on: http://gerrit.usersys.redhat.com/1021
Reviewed-by: Dan Kenigsberg <danken@redhat.com>
Tested-by: Dan Kenigsberg <danken@redhat.com>

Refactor dispatcher unicode handling

Change-Id: Id007710654e5ddbb41c269f624646159fbbc904e
Reviewed-on: http://gerrit.usersys.redhat.com/1020
Reviewed-by: Dan Kenigsberg <danken@redhat.com>
Tested-by: Dan Kenigsberg <danken@redhat.com>

Made the logging mechanism more generic

 - modified storage connection verification to have better errors
 - pulled out logging logic from dispatcher to a generic decorator

Change-Id: I38b4a16803b03043144acf62e9fb65c3b1b0a3c5
Reviewed-by: Dan Kenigsberg <danken@redhat.com>
Tested-by: Dan Kenigsberg <danken@redhat.com>

BZ#745390 always report product=RHEV to guest bios

Laziness is dangerous; patches should be checked.

Change-Id: I5dee9473c7570134cb8ade66efa07fe223d59b55
Reviewed-on: http://gerrit.usersys.redhat.com/1055
Reviewed-by: Dan Kenigsberg <danken@redhat.com>
Tested-by: Dan Kenigsberg <danken@redhat.com>

Removing references to SDF.

Change-Id: I87ca502ebbb1c6384a8fe3d5f02edf3ed6d084f5
Reviewed-on: http://gerrit.usersys.redhat.com/995
Tested-by: Dan Kenigsberg <danken@redhat.com>
Reviewed-by: Dan Kenigsberg <danken@redhat.com>

Remove sdf.py

Change-Id: I19d178e3b37fa119dc8a44815bb5055535c95241
Reviewed-on: http://gerrit.usersys.redhat.com/644
Reviewed-by: Dan Kenigsberg <danken@redhat.com>
Tested-by: Dan Kenigsberg <danken@redhat.com>

Simplify SDF.produce() for removing sdf.py.

produce() will return a domain or raise StorageDomainDoesNotExist.
No need for this check.

Change-Id: I442c368a1266855b324eab96e4a7b02634984834
Reviewed-on: http://gerrit.usersys.redhat.com/643
Reviewed-by: Dan Kenigsberg <danken@redhat.com>
Tested-by: Dan Kenigsberg <danken@redhat.com>

Remove SDF.create() for removing sdf.py

Change-Id: I31b64def7aea87d4d3a94950abc8e96fbe3c4b8a
Reviewed-on: http://gerrit.usersys.redhat.com/642
Reviewed-by: Dan Kenigsberg <danken@redhat.com>
Tested-by: Dan Kenigsberg <danken@redhat.com>

Simplify recycle function for removing the sdf.py

Change-Id: I9cf3106f0fb5d5cd067907d48ff02fe50493a66c
Reviewed-by: Dan Kenigsberg <danken@redhat.com>
Tested-by: Dan Kenigsberg <danken@redhat.com>

Change format signature in order to simplify SD recycle.

Needed for remove the sdf.py

Change-Id: Iea1e5735563087195181c625e57b8ba8e75d3441
Reviewed-by: Dan Kenigsberg <danken@redhat.com>
Tested-by: Dan Kenigsberg <danken@redhat.com>

BZ#741203 Use cached node id if SPM is active

Do not invalidate the metadata cache if the node is the SPM and it
is active. This will prevent the getStoragePoolInfo call to be stuck
when the domain is unreachable and at the same time won't break the
fix for BZ#733909.

In the patch:
* add the attribute isSPM to the StoragePool class
* the SPM never requires to invalidate the cache

Change-Id: Iee15c4b35a415e9ea460b1ce6a4afa14d147ed8d
Reviewed-on: http://gerrit.usersys.redhat.com/1004
Reviewed-by: Dan Kenigsberg <danken@redhat.com>
Tested-by: Dan Kenigsberg <danken@redhat.com>

BZ#741203 Don't return a value in domain selftest

The domain selftest doesn't need to return a boolean value, if the check
fails an exception is raised.

Change-Id: I551970c1f426f2b28ad58e49310338de6b40a29f
Reviewed-on: http://gerrit.usersys.redhat.com/1012
Reviewed-by: Dan Kenigsberg <danken@redhat.com>
Tested-by: Dan Kenigsberg <danken@redhat.com>

BZ#746766 deployUtil: set proper selinux context to .ssh/authorized_hosts

Change-Id: I6833b991f89033b0b6ab0c85fb6f8df49809ea26
Reviewed-on: http://gerrit.usersys.redhat.com/1045
Tested-by: Dan Kenigsberg <danken@redhat.com>
Reviewed-by: Dan Kenigsberg <danken@redhat.com>

BZ#745390 always report product=RHEV to guest bios

Two years after insisting the ueber-importance of differentiating RHEL
hosts from RHEV ones, RHN/Satellite suddenly needs us to always report
"RHEV-H" to the guest, or else the guest would not be recognized as
virtual by them.

Disagreing and committing.

Change-Id: Ibcee345c6960ffb35f517a7dd77d5919c4101afa
Reviewed-on: http://gerrit.usersys.redhat.com/1039
Reviewed-by: Dan Kenigsberg <danken@redhat.com>
Tested-by: Dan Kenigsberg <danken@redhat.com>

BZ#746975 - Allow deactivate storage domain if not reachable

Change-Id: Id547f6bdb1a8e5c41340981c9453a98d099d713e
Reviewed-on: http://gerrit.usersys.redhat.com/1044
Tested-by: Igor Lvovsky <ilvovsky@redhat.com>
Reviewed-by: Dan Kenigsberg <danken@redhat.com>

BZ#741658 Move VM to Down if libvirt reports VIR_ERR_NO_DOMAIN

Change-Id: Id6262fb2504534c56fdec246937423d629f92e13
Reviewed-on: http://gerrit.usersys.redhat.com/984
Reviewed-by: Dan Kenigsberg <danken@redhat.com>
Tested-by: Dan Kenigsberg <danken@redhat.com>

Correct the vdsmd init script path

Change-Id: I0b5ab2341675d66976ce9879bf6a17f4bd881ac7
Reviewed-on: http://gerrit.usersys.redhat.com/1038
Tested-by: Dan Kenigsberg <danken@redhat.com>
Reviewed-by: Dan Kenigsberg <danken@redhat.com>

Remove unused symlinks

Change-Id: I79d713912ba1b9ee980610bd0870af95f8e2b457
Reviewed-on: http://gerrit.usersys.redhat.com/1026
Reviewed-by: Dan Kenigsberg <danken@redhat.com>
Tested-by: Dan Kenigsberg <danken@redhat.com>

Minor fixes for distcheck

In this patch:
* don't distribute files generated by the substitutions
* use srcdir as path for substitutions
* add uninstall-local for the storage subdirectory

Change-Id: I7b05aeb82ce2caf4ecc9d860ce1fbc436f5d465d
Reviewed-on: http://gerrit.usersys.redhat.com/1018
Tested-by: Dan Kenigsberg <danken@redhat.com>
Reviewed-by: Dan Kenigsberg <danken@redhat.com>

BZ#743549 - Allow concurrency in task manager

Change-Id: I7b04a8b22ac40e39e89bbbee56309c015586fc1f
Reviewed-on: http://gerrit.usersys.redhat.com/997
Reviewed-by: Dan Kenigsberg <danken@redhat.com>
Tested-by: Rami Vaknin <rvaknin@redhat.com>

BZ#743887 Close lifeline pipe when oop is killed

Change-Id: Ife49c57106642b5d5c823867af594d5e4c5bc470
Reviewed-on: http://gerrit.usersys.redhat.com/1024
Reviewed-by: Dan Kenigsberg <danken@redhat.com>
Reviewed-by: Ayal Baron
Tested-by: Dan Kenigsberg <danken@redhat.com>

BZ#744549 getVdsCaps: survive multiple redhat-release rpms

caps._getKeyPackages() already handles this case. Too bad this was
forgotten in osversion().

Change-Id: I2ce9b706817d0fe49a043970d2b96b8c674f32c8
Reviewed-on: http://gerrit.usersys.redhat.com/999
Reviewed-by: Dan Kenigsberg <danken@redhat.com>
Tested-by: Dan Kenigsberg <danken@redhat.com>

Include vdscli.py in the vdsm rpm

Since we moved to kaxmlrpclib we require vdscli in vm.py which is part
of vdsm.

Change-Id: Iff31e3404424c25fa95a3f581ec114a20e00a4c4
Reviewed-on: http://gerrit.usersys.redhat.com/1019
Reviewed-by: Dan Kenigsberg <danken@redhat.com>
Tested-by: Dan Kenigsberg <danken@redhat.com>

BZ#742764 - Don't change template's type during createVolume rollback

Change-Id: Id490037578ba4acde1514c6f80713d8aec1831d8
Reviewed-on: http://gerrit.usersys.redhat.com/1025
Tested-by: Igor Lvovsky <ilvovsky@redhat.com>
Reviewed-by: Ayal Baron
Reviewed-by: Dan Kenigsberg <danken@redhat.com>

Need to also catch AuthenticationError in case connecting to the SuperVdsm fails.

Change-Id: I101c5108593425d8f7e47b927bc49ec0b75b1af8
Reviewed-on: http://gerrit.usersys.redhat.com/1016
Tested-by: Yotam Oron <yoron@redhat.com>
Reviewed-by: Dan Kenigsberg <danken@redhat.com>

BZ#732980 Initialize the block sizes metadata

Change-Id: I0baa4addabc31389ea1656099f7c9d2d70cd85a0
Reviewed-on: http://gerrit.usersys.redhat.com/992
Reviewed-by: Dan Kenigsberg <danken@redhat.com>
Tested-by: Dan Kenigsberg <danken@redhat.com>

BZ#736114 bootstrap: override iptables if requested

vds_bootstrap script gets a new optional argument -f <filename>.
vds_installer will use the argument to pass the local location of an iptables
rule file to override /etc/sysconfig/iptables.

Change-Id: I6a2e7d0512a2c9fc691135bf42b03fdcb3ee088a
Reviewed-on: http://gerrit.usersys.redhat.com/945
Tested-by: David Naori <dnaori@redhat.com>
Reviewed-by: Dan Kenigsberg <danken@redhat.com>

Related to BZ#743272 - dirList IS and WAS dirList.

Change-Id: I0fbcdb4076a5c16fd74754350464f382636ff419
Reviewed-on: http://gerrit.usersys.redhat.com/993
Reviewed-by: Ayal Baron
Tested-by: Dan Kenigsberg <danken@redhat.com>
Reviewed-by: Dan Kenigsberg <danken@redhat.com>

Related to BZ#738120 - Logging in OOP - drop identical messages.

Somehow we get every log 3 times in the queue.
In the future we will be need to find out the real reason
for such behavior and fix it. For now just drop identical messages.

Change-Id: I0b48d6111c87eaa8253984e378fe027cc4639510
Reviewed-on: http://gerrit.usersys.redhat.com/1008
Reviewed-by: Dan Kenigsberg <danken@redhat.com>
Tested-by: Dan Kenigsberg <danken@redhat.com>

Related to BZ#738120 - Logging in OOP.

Now we can use logging in OOP operations.

Change-Id: Iabe8abf31f46469e3c6fdac670f0ff03fe9cefe8
Reviewed-on: http://gerrit.usersys.redhat.com/985
Reviewed-by: Dan Kenigsberg <danken@redhat.com>
Tested-by: Dan Kenigsberg <danken@redhat.com>

Add partial vms list support to the regular 'list' query.

Usage: vdsClient 0 list [table/ids/long] [vms:vmId1,vmId2]

Change-Id: I5051c42ca9e4ce706eff3cfec0de04cd76b2b95f
Reviewed-on: http://gerrit.usersys.redhat.com/956
Reviewed-by: Dan Kenigsberg <danken@redhat.com>
Tested-by: Igor Lvovsky <ilvovsky@redhat.com>

BZ#743828 allocate only 32MB vram for multiple qxl monitors

Change-Id: I3d0f32a931cc0e87185e6561e94c0dd57e422a9f
Reviewed-on: http://gerrit.usersys.redhat.com/1005
Reviewed-by: Dan Kenigsberg <danken@redhat.com>
Tested-by: Dan Kenigsberg <danken@redhat.com>

BZ#740158 netinfo.speed(): never report -1 speed

Related to Bug 709674: sysfs may report -1 speed (in unsigned 16- and
32-bit integers) for disabled nics. Vdsm should report their speed as 0.

Change-Id: I309bc4940405ebfe73fdcbe36f83db33a426fee5
Reviewed-on: http://gerrit.usersys.redhat.com/1001
Reviewed-by: Yotam Oron <yoron@redhat.com>
Reviewed-by: Dan Kenigsberg <danken@redhat.com>
Tested-by: Dan Kenigsberg <danken@redhat.com>

BZ#720385 - Report disk latency (read , write & flush ) for each storage device.

Change-Id: Id0d3d4869bcf35ed4d2ff72a5e1ef7aa21fe3455
Reviewed-on: http://gerrit.usersys.redhat.com/935
Reviewed-by: Dan Kenigsberg <danken@redhat.com>
Tested-by: Dan Kenigsberg <danken@redhat.com>

BZ#741856 delete rhevm cert before conf_update

Delete CA certificate before updating vdsm-reg.conf, to avoid certificate
collision in the case of later re-configuration with no fingerprint verification

Change-Id: I8bc1dc1fb7fa2ffc0bcb1f6389b34a9ee82b4197
Reviewed-on: http://gerrit.usersys.redhat.com/987
Tested-by: Yotam Oron <yoron@redhat.com>
Reviewed-by: Ayal Baron
Reviewed-by: Dan Kenigsberg <danken@redhat.com>

BZ#740887 Tune cache dirty ratio

Tuning the dirty_ratio and dirty_background_ratio kernel parameters
increases I/O throughput from the guests, improves fairness between
the guests and reduces the ability of a buffered writer to starve
guests.

Change-Id: Ibf5c8e4c0637c60092b89fba103b96b37bdafaa0
Reviewed-on: http://gerrit.usersys.redhat.com/970
Reviewed-by: Dan Kenigsberg <danken@redhat.com>
Tested-by: Dan Kenigsberg <danken@redhat.com>

BZ#732765 - Drop disktotal/diskfree parameters for unreachable domains

Change-Id: Ie359e5fb70cd5dc2e59f8d9e166fb02ba3307273
Reviewed-on: http://gerrit.usersys.redhat.com/1003
Tested-by: Igor Lvovsky <ilvovsky@redhat.com>
Reviewed-by: Ayal Baron
Reviewed-by: Dan Kenigsberg <danken@redhat.com>

BZ#737722 Do not pass cmdline args to grub in RHEV-H

cmdline args are only needed in installation time,
oVirt has a builtin functionallity to not pass the args
to grub

Change-Id: I3347e60ecbf88f3c8e198d36104dfc55ebdee288
Reviewed-on: http://gerrit.usersys.redhat.com/988
Reviewed-by: Dan Kenigsberg <danken@redhat.com>
Tested-by: Dan Kenigsberg <danken@redhat.com>

BZ#743122 libvirtvm.create: honor tdf param

Some guests (mostly Windows) need the time-drift-fix to keep up their
clock.

v2
- Invert default, as KVM's Dor says it is safe enough for all guests.

Change-Id: I0dd35ebfa7bc327455cf0b2739fbdaa16c752edf
Reviewed-on: http://gerrit.usersys.redhat.com/994
Tested-by: Dan Kenigsberg <danken@redhat.com>
Reviewed-by: Yotam Oron <yoron@redhat.com>
Reviewed-by: Dan Kenigsberg <danken@redhat.com>

Don't pollute the root directory during installation

Change-Id: I6c8e4f4253e03f6fa6fe249f29eeac1ea4619817
Reviewed-on: http://gerrit.usersys.redhat.com/996
Tested-by: Dan Kenigsberg <danken@redhat.com>
Reviewed-by: Dan Kenigsberg <danken@redhat.com>

BZ#736301 - Don't fail lvm.getVGbyUUID() if unreadable VGs found.

Change-Id: Ief82082ebb3cfb6731dc9c7905f2844c60a7f425
Reviewed-on: http://gerrit.usersys.redhat.com/930
Reviewed-by: Dan Kenigsberg <danken@redhat.com>
Tested-by: Dan Kenigsberg <danken@redhat.com>

Remove unused TaskResource class

Change-Id: Ib487ff8661c771b765afed1f97327ef88c798745
Reviewed-on: http://gerrit.usersys.redhat.com/980
Tested-by: Dan Kenigsberg <danken@redhat.com>
Reviewed-by: Dan Kenigsberg <danken@redhat.com>

Correct typos for 'successfully'

Change-Id: I8e5bfc5911f3a981e20bf721fb348c0bcd4e9d5d
Reviewed-on: http://gerrit.usersys.redhat.com/979
Reviewed-by: Dan Kenigsberg <danken@redhat.com>
Tested-by: Dan Kenigsberg <danken@redhat.com>

BZ#742267 server_connection: checking access is enough

There is no need to verify that a directory/mountpoint is owned by
vdsm:kvm. All we need is vdsm accessibility. Ownership requirement was
added in downstream commit 9fb05359 for local storage, and later
(http://gerrit.usersys.redhat.com/776) extended to all storage types.

Change-Id: I799374f13cc9d5fe193d76479f738e0a5fa02836
Reviewed-on: http://gerrit.usersys.redhat.com/983
Reviewed-by: Igor Lvovsky <ilvovsky@redhat.com>
Tested-by: David Naori <dnaori@redhat.com>
Reviewed-by: Dan Kenigsberg <danken@redhat.com>

BZ#735621 fileSD: Fix remotePath in SD metadata (V3)

Changes since V2:
 - Rename mountToRemotePath() to getRealPath() and define it in each child
   class.

Changes since V1:
 - Derive the remotePath from self.mountpoint instead of using the metadata

The current method for gathering a LOCALFS Storage Domain's remotePath
property does not work because these domains are connected with a symlink,
not a mount.  Fix up the current code so that it handles links and
mountpoints.

Change-Id: If50379ad0685fad35025ab142c776f2f8a3c5e30
Signed-off-by: Adam Litke <agl@us.ibm.com>

Related to BZ#736422: Fixed a typo in command's name.

The RHEV-Agent send a "session-lock" notification but vdsm
expected to receive a "session-locked".

Change-Id: Icef91042c6d8719d649249bada8c7df332f41db3
Reviewed-on: http://gerrit.usersys.redhat.com/973
Reviewed-by: Dan Kenigsberg <danken@redhat.com>
Tested-by: Jakub Libosvar <libosvar@redhat.com>

BZ#716573 make vdsm's `ulimit -n` configurable

Change-Id: Ia7e8282770650a691e8d388c62e64111df40b6b2
Reviewed-on: http://gerrit.usersys.redhat.com/948
Reviewed-by: Dan Kenigsberg <danken@redhat.com>
Tested-by: Dan Kenigsberg <danken@redhat.com>

BZ#741190 never add a second "vdsm" stanza to logrotate.d/libvirtd

Change-Id: I6603ae0562481199854f6986931efdb30d43d9c8
Reviewed-on: http://gerrit.usersys.redhat.com/971
Reviewed-by: Dan Kenigsberg <danken@redhat.com>
Tested-by: Dan Kenigsberg <danken@redhat.com>

BZ#732980 Devices block size must be 512

Change-Id: If303fac85a29bed7b989b0b6302fc4c83075f500
Reviewed-on: http://gerrit.usersys.redhat.com/855
Reviewed-by: Dan Kenigsberg <danken@redhat.com>
Tested-by: Dan Kenigsberg <danken@redhat.com>

Don't use a list to calculate VG size

Change-Id: Ief2a3e69f94c99b11c9027901fc5a34577c3452f
Reviewed-on: http://gerrit.usersys.redhat.com/967
Reviewed-by: Dan Kenigsberg <danken@redhat.com>
Tested-by: Dan Kenigsberg <danken@redhat.com>

BZ#732980 Check block size on VG commands

Change-Id: Ie106491328f12f650b6bb234a92238d6c0aa70c8
Reviewed-by: Ayal Baron
Reviewed-by: Saggi Mizrahi <smizrahi@redhat.com>
Reviewed-by: Dan Kenigsberg <danken@redhat.com>
Tested-by: Dan Kenigsberg <danken@redhat.com>

BZ#732980 MD tag must be in blocks unit

This patch fixes the unit used for the MD tag (blocks instead of
bytes) and completes the support for the new MS tag which holds
the size (in blocks) of the metadata.

Change-Id: Id268735f60a289f9e8f12a7d7bb3180bf3398c54
Reviewed-on: http://gerrit.usersys.redhat.com/891
Reviewed-by: Ayal Baron
Reviewed-by: Saggi Mizrahi <smizrahi@redhat.com>
Tested-by: Dan Kenigsberg <danken@redhat.com>
Reviewed-by: Dan Kenigsberg <danken@redhat.com>

BZ#735725 - Insert a configuration line if not exists

- Simple configuration line insertion if missing

Change-Id: Ica24a08b66c076e87753cfb198f28822f8ecc044
Reviewed-on: http://gerrit.usersys.redhat.com/972
Reviewed-by: Dan Kenigsberg <danken@redhat.com>
Tested-by: Haim Ateya <hateya@redhat.com>

Replace m2crypto: re-add kaxmlrpclib

Change-Id: I4f03ebfe4f9ab8e153fb38cbbacf0f15035b7ccf
Reviewed-by: Dan Kenigsberg <danken@redhat.com>
Tested-by: Dan Kenigsberg <danken@redhat.com>

SecureXMLRPC: work with python 2.7, too.

Change-Id: Id09274ed848632aced7f66bf5322f2f5204ce89e
Reviewed-by: Dan Kenigsberg <danken@redhat.com>
Tested-by: Dan Kenigsberg <danken@redhat.com>

Replace m2crypto: re-add timeout

Change-Id: I8003bfc7eba33de53588acc8a2925bcbcea69fe0
Reviewed-by: Dan Kenigsberg <danken@redhat.com>
Tested-by: Dan Kenigsberg <danken@redhat.com>

read thp state from its upstream location first

Change-Id: I174d1d86a89af405b6bd1f04401e07934ed940d6
Reviewed-by: Dan Kenigsberg <danken@redhat.com>
Tested-by: Dan Kenigsberg <danken@redhat.com>

make /etc/init.d/vdsmd workable under F15

Change-Id: I50e8859799b8a86d47a50b9fbb068b603a42d49b
Reviewed-by: Dan Kenigsberg <danken@redhat.com>
Tested-by: Dan Kenigsberg <danken@redhat.com>

Replace m2crypto with python's standard ssl

Change-Id: Ie2162f72c914dc6c7832ca2951bbf151ed058859
Reviewed-by: Dan Kenigsberg <danken@redhat.com>
Tested-by: Dan Kenigsberg <danken@redhat.com>