WIP

Change-Id: I6e65687fbfa6562c3a8715341a75e27731dd75f9

pidl-wireshark: pretend to use size variable passed a parameter it the type is a blob

That's because data blob don't use the size information from the
parameter instead it reads it on the wire.

Change-Id: I5b4631b6ac93a3c90f9f56ef9fb72ea96c34c96d
Signed-off-by: Matthieu Patou <mat@matws.net>

git add -f bin/default/librpc/gen_ndr/*.{c,h,ndr,pidl} bin/default/source*/librpc/gen_ndr/*.{c,h,ndr,pidl}

Change-Id: I54d98b26f4543e80691a1505eaede87846189915

idl: reduce the number of IDL that are not handled by PIDL for wireshark generator

Change-Id: I5b0a3da2b6962839918036557ceb77756e3c64c4

pidl-wireshark: register dissector for type uid_t and gid_t

Change-Id: I101e803d298251bf8acd9bea4275233079a8c35b

pidl-wireshark: subcontext stuff IS IT NEEDED ?

Change-Id: I06dd50099cf6ac28619975f57b86388abda4e0b5

pidl-wireshark: reduce the number of false positive about un-usued dissectors or ETT

Change-Id: Ie9cea6367db6ad44f14bd018a1b59a32b20844b8

pidl-wireshark: if param is specified store the value of the dissected bitmap

Change-Id: I7da4ac93370c46373bd6e476860bccc09a131e4a

pidl-wireshark: handle properly bitmap of different size

Change-Id: I0aec8ad03d126d57e98402dd12937e4065675c84

pidl-wireshark: do not generate bitmap dissection function if the bitmap has no elements

Change-Id: I97adb0dac88652bd1e4a996bf053528e1a6719b2

pidl-wireshark: adapt to changes in wireshark where dcerpc_info is now passed to each functions

Change-Id: Icb7e9c9df9e0497062ffb4742c25bd86e85717ba

pidl-wireshark: handle relative pointers

Change-Id: I4efa8e622165e1b4c417cccae74d3e2aa64c0056

pidl-wireshark: handle cases where the context size is not not on the wire but known in advance

Signed-off-by: Matthieu Patou <mat@matws.net>

pidl-wireshark: handle case where array_size is an expression and not just a variable

For instance [charset(UTF16)] uint16 file_name[FRSRPC_MAX_PATH+1];

Change-Id: Ifcb1e65777d1d6dd85af649be3d61b278f96f84b
Signed-off-by: Matthieu Patou <mat@matws.net>

pidl-wireshark: generate an adaptation function when dissecting pointers to type that need parameters

If in order to dissect a variable I need parameters (switch_type,
size_is ...) and the variable is actually a pointer we will generate an
adaptation function that will have the expected signature for the
dissect_ndr_xxx_pointer but that will be able to extract data from
dcv->private data to inject them in the real dissector that needs
parameters

Change-Id: I7ca2af53ab7ffadb5c0103e554933af7a8611d18
Signed-off-by: Matthieu Patou <mat@matws.net>

pidl-wireshark: do not generate dissection code for bitmap if bitmap is emtpy

Change-Id: If9346fb2fcbec2dae093002b8909c236817c4ffb
Signed-off-by: Matthieu Patou <mat@matws.net>

HACK pidl-wireshark: align pointer

Change-Id: Ibcd7c3c3a0cf63670f661c22c1f6a3779ccba978
Signed-off-by: Matthieu Patou <mat@matws.net>

pidl-wireshark: remove leading space in the field name

Change-Id: Icd1617eb4abe0895794c8eee34db4a1be3ef1a5f
Signed-off-by: Matthieu Patou <mat@matws.net>

pidl-wireshark: do the registration of the type before parsing as some element of the current type might be a pointer to this type too

Ie. to deal with things like that:
typedef [public,noprint] struct {
drsuapi_DsReplicaObjectListItemEx *next_object;
drsuapi_DsReplicaObject object;
boolean32 is_nc_prefix;
GUID *parent_object_guid;
drsuapi_DsReplicaMetaDataCtr
*meta_data_ctr;
}
drsuapi_DsReplicaObjectListItemEx;

Signed-off-by: Matthieu Patou <mat@matws.net>

pidl-wireshark: don't generate body of the function is there is no elements in it

Change-Id: Iac26f5987d6f4102b09ee429d1b481070a663b99
Signed-off-by: Matthieu Patou <mat@matws.net>

pidl-wireshark: Create a separate hf for the switch variable of union

Before we were using the HF of the union itself but it was missing the
FT_UINTxx which cause a problem

Signed-off-by: Matthieu Patou <mat@matws.net>

pidl-wireshark: use PIDL_dissect instead of dissect_ndr: it's more readable

Signed-off-by: Matthieu Patou <mat@matws.net>

pidl-wireshark: Add TRY/CATCH in subcontext so that if the parsing fail we still restore the NDR flags

Signed-off-by: Matthieu Patou <mat@matws.net>

pidl-wireshark: extract the loop for element from the Struct() function

Change-Id: I569284fecd2f79b0c8b9f432e63df847e8f495aa

pidl-wireshark: Extract variable name of switch_is and subcontext_size

This allow to handle cases like
[switch_is(wDataLength?id:DSPROPERTY_ZONE_EMPTY)]     dnsPropertyData data;

Otherwise we were trying to declare variable called
wDataLength?id:DSPROPERTY_ZONE_EMPTY and it was not working very well

Change-Id: I82460228de354cdd00630ccc670ac565d7a9ccc8
Signed-off-by: Matthieu Patou <mat@matws.net>

pidl-wireshark: if the header size has a 0x then it's the size of the subcontext not the size of the variable for length of subcontext

Signed-off-by: Matthieu Patou <mat@matws.net>

pidl-wireshark: try to dissect the subcontext size only if the HEADER_SIZE is not null

Signed-off-by: Matthieu Patou <mat@matws.net>

pidl-wireshark: Handle subcontext that have the size specified by an external variable

Signed-off-by: Matthieu Patou <mat@matws.net>

pidl-wireshark: add definition for IPV4/IPV6 types

Change-Id: I35bdf98ad7f6de38645531eb1b945297cfb704f2
Signed-off-by: Matthieu Patou <mat@matws.net>

pidl-wireshark: Add capability to pass more than one parameter

In some construction we might need to pass 2 or more parameters.
One for the switch value and one for the size of the subcontext (for
instance)

Change-Id: Ia68e2e70fe84e543971b9ebf313a7ecaf3def1ca
Signed-off-by: Matthieu Patou <mat@matws.net>

pidl-wireshark: Do not generate dissection for WERROR field if the function contains that have BytePipe elements

Change-Id: I80931ea246d46ab37e14fcd0ecaf5157387bc205
Signed-off-by: Matthieu Patou <mat@matws.net>

pidl-wireshark: generate the dissector for DCE-RPC pipe

Change-Id: I6d57518f944b21fb1c426df09bf34fdcc51bb36e
Signed-off-by: Matthieu Patou <mat@matws.net>

pidl-wireshark: Do not add the signature of the dissector if it was manually defined

The signature can be different and as manual dissectors are at the top
of the file they will be defined before being used

Change-Id: I8d487c3cda2b6d7249706bc842c52227fb23f5a5
Signed-off-by: Matthieu Patou <mat@matws.net>

pidl-wireshark: clean case name as they can have no valid char for variable names

For instance:
[case(1|(DRSUAPI_COMPRESSION_TYPE_MSZIP<<16))]16drsuapi_DsGetNCChangesMSZIPCtr1
mszip1;

Signed-off-by: Matthieu Patou <mat@matws.net>

pidl-wireshark: Add code for dissecting inline arrays

If the array contains strings, display as strings instead of array of
bytes/dword/...

Change-Id: I1a24886b63271a94c7fb26ef3af5f4f7afe1f9e9
Signed-off-by: Matthieu Patou <mat@matws.net>

pidl-wireshark: if the element is used as a switch store the value

We might need this value later on

Change-Id: I7c3c69cde3a05f623007911c7e5f6756a3c60ee3
Signed-off-by: Matthieu Patou <mat@matws.net>

pidl-wireshark: Dissect strings inside array as string rather as bytes

Change-Id: Icf9b994cdb0ec6d6c61049e8e7235c50e3ff9b6d

pidl-wireshark: Display Array as block instead of bytes by bytes

Change-Id: Ie6062990f0fa5389772cc3839b7dd8817a1617a3

Regenerate idl stuff with git add -f bin/default/librpc/gen_ndr/*.{c,h,ndr,pidl} bin/default/source*/librpc/gen_ndr/*.{c,h,ndr,pidl}

Change-Id: Ica1cc1243914b49b5ddafd1ed3557b3f0488df5b

Add generated files git add -f bin/default/librpc/gen_ndr/*.{c,h,ndr,pidl} bin/default/source*/librpc/gen_ndr/*.{c,h,ndr,pidl}

Change-Id: I287ce654d3c1fc2657ce057c0d87aeda02cb02d5
Signed-off-by: Matthieu Patou <mat@matws.net>

HACK librpc/idl: --ws-parser

Change-Id: I0326d4f3fc416d8d722cf5889732b2f75e3ac144

HACK: pidl: always dump idl and ndr tree

librpc: add idl for DFS-R (frsrtans) blobs

Change-Id: Ie9653abcea46f7e347ff595064f17d20bcc93ca1
Signed-off-by: Matthieu Patou <mat@matws.net>

WIP DEBUG: use Data::Dumper

Change-Id: Icfb7efda658e71ae22678466aab0717675219408

Remove [u]int[XX] defines from source3/include/includes.h

Signed-off-by: Richard Sharpe <rsharpe@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
Autobuild-User(master): Jeremy Allison <jra@samba.org>
Autobuild-Date(master): Mon May 18 05:24:22 CEST 2015 on sn-devel-104

Use tevent_req_poll_ntstatus

Kill 41 lines ..

Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>

Convert the few instances of int32 there were to int32_t.

Signed-off-by: Richard Sharpe <rsharpe@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
Autobuild-User(master): Jeremy Allison <jra@samba.org>
Autobuild-Date(master): Sat May 16 03:09:12 CEST 2015 on sn-devel-104

pidl:Samba3/ServerNDR: normalize {RETURN_TYPE} by using mapTypeName()

Signed-off-by: Stefan Metzmacher <metze@samba.org>
Reviewed-by: Richard Sharpe <realrichardsharpe@gmail.com>
Autobuild-User(master): Jeremy Allison <jra@samba.org>
Autobuild-Date(master): Fri May 15 22:19:38 CEST 2015 on sn-devel-104

Convert uint64 to uint64_t

We seemed to have very few uses of that.

Signed-off-by: Richard Sharpe <rsharpe@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>

Covert all uint32/16/8 to _t in source3/rpc_server.

This can be committed regardless of the state of the PIDL changes.

Signed-off-by: Richard Sharpe <rsharpe@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>

s3: IPv6 enabled DNS connections for ADS client

This patch makes DNS client connections protocol independent.
For example DNS updates. This makes IPv6-only clients possible.

Signed-off-by: David Holder <david.holder@erion.co.uk>
Reviewed-by: Jeremy Allison <jra@samba.org>
Reviewed-by: Ralph Böhme <rb@sernet.de>

Convert all uint32/16/8 to _t in a grab-bag of remaining files.

I still need to fix the rpc stuff, but we are almost there.

Signed-off-by: Richard Sharpe <rsharpe@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
Autobuild-User(master): Jeremy Allison <jra@samba.org>
Autobuild-Date(master): Thu May 14 22:16:56 CEST 2015 on sn-devel-104

Convert all uses of uint32/16/8 to _t in source3/groupdb.

Signed-off-by: Richard Sharpe <rsharpe@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>

namequery: Fix a missing TALLOC_FREE

Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Ira Cooper <ira@samba.org>
Autobuild-User(master): Volker Lendecke <vl@samba.org>
Autobuild-Date(master): Thu May 14 16:49:39 CEST 2015 on sn-devel-104

Fix the uint32/16/8 stuff in two proto.h files I missed.

Signed-off-by: Richard Sharpe <rsharpe@gmail.com>
Reviewed-by: Jeremy Allison <jra@samba.org>
Autobuild-User(master): Jeremy Allison <jra@samba.org>
Autobuild-Date(master): Wed May 13 22:01:14 CEST 2015 on sn-devel-104

Convert all uses of uint32/16/8 to _t in source3/auth.

Signed-off-by: Richard Sharpe <rsharpe@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>

Convert uint32/16/8 to _t in source3/libnet.

Signed-off-by: Richard Sharpe <rsharpe@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>

Convert all uint32/16/8 to _t in source3/client.

Signed-off-by: Richard Sharpe <rsharpe@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>

Convert a couple of uint32s to _t that I missed in source3/utils.

Signed-off-by: Richard Sharpe <rsharpe@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>

Convert all uint32/16/8 to _t in source3/libsmb.

Signed-off-by: Richard Sharpe <rsharpe@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>

Convert all uint32/16/8 to _t in source3/lib.

Signed-off-by: Richard Sharpe <rsharpe@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>

Convert all uint32/16/8 to _t in source3/locking.

Signed-off-by: Richard Sharpe <rsharpe@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>

namequery: namequery: suppress bogus warning "resolve_name: unknown name switch type"

Based on a patch by Uri Simchoni <urisimchoni@gmail.com>

Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>

ctdb-doc: Document externally managed public IP addresses

For use with DisableIPFailover=1.

Pair-programmed-with: Amitay Isaacs <amitay@gmail.com>
Signed-off-by: Martin Schwenke <martin@meltin.net>
Reviewed-by: Amitay Isaacs <amitay@gmail.com>
Autobuild-User(master): Amitay Isaacs <amitay@samba.org>
Autobuild-Date(master): Wed May 13 09:27:36 CEST 2015 on sn-devel-104

ctdb-tests: Add tickle test for external public address handling

Signed-off-by: Martin Schwenke <martin@meltin.net>
Reviewed-by: Amitay Isaacs <amitay@gmail.com>

ctdb-scripts: New eventscript 10.external

This is an alternative to 10.interface and is installed as disabled by
default.  It should only be used with DisableIPFailover=yes and when
IP failover is being handled externally.  In this mode CTDB can be
informed of public IP address movements using "ctdb moveip".

During the "startup" event, this eventscript currently finds any
public IP addresses configured in $CTDB_PUBLIC_ADDRESSES and tells
CTDB which node they are on using "ctdb moveip".  This allows CTDB to
send ARPs and tickle-ACKs.

Pair-programmed-with: Amitay Isaacs <amitay@gmail.com>
Signed-off-by: Martin Schwenke <martin@meltin.net>
Reviewed-by: Amitay Isaacs <amitay@gmail.com>

ctdb-daemon: Move release of all IPs to startup

This means that DisableIPFailover will be set if it should be.

Signed-off-by: Martin Schwenke <martin@meltin.net>
Reviewed-by: Amitay Isaacs <amitay@gmail.com>

ctdb-daemon: Never release all IPs when DisableIPFailover is set

If DisableIPFailover is set then something else may be managing public
IP addresses so CTDB should leave them alone.

Pair-programmed-with: Amitay Isaacs <amitay@gmail.com>
Signed-off-by: Martin Schwenke <martin@meltin.net>
Reviewed-by: Amitay Isaacs <amitay@gmail.com>

ctdb-scripts: Drop all public IP addresses from 10.interface

00.ctdb should not know about public IP addresses.

Move related tests to operate on 10.interface.

Pair-programmed-with: Amitay Isaacs <amitay@gmail.com>
Signed-off-by: Martin Schwenke <martin@meltin.net>
Reviewed-by: Amitay Isaacs <amitay@gmail.com>

ctdb-daemon: Don't update IP tree if DisableIPFailover is set

There won't be an IP tree.  It is only ever initialised during a
takeover run.

The alternate to this would be to avoid sending
CTDB_SRVID_RECD_UPDATE_IP in "ctdb moveip".  This logic is probably
best kept out of the CLI tool.

Pair-programmed-with: Amitay Isaacs <amitay@gmail.com>
Signed-off-by: Martin Schwenke <martin@meltin.net>
Reviewed-by: Amitay Isaacs <amitay@gmail.com>

ctdb-daemon: Mark interfaces as "up" by default

The potential for public IP addresses to shuffle around during node
initialisation disappeared a while ago because IP addresses can only
be assigned to a node that is in CTDB_RUNSTATE_RUNNING.  This means
that interfaces might as well just be initialised as "up".  If any
interfaces are actually "down" then this will be rectified by the
"startup" event in 10.interfaces.

Signed-off-by: Martin Schwenke <martin@meltin.net>
Reviewed-by: Amitay Isaacs <amitay@gmail.com>

ctdb-daemon: Skip "IP on interface" checks if DisableIPFailover is set

To support external failover of IP addresses if DisableIPFailover is
set.  CTDB's idea of IP address assignment can be manipulated using
"ctdb moveip".  Checking if the IP address is already held breaks
this in several places.

Pair-programmed-with: Amitay Isaacs <amitay@gmail.com>
Signed-off-by: Martin Schwenke <martin@meltin.net>
Reviewed-by: Amitay Isaacs <amitay@gmail.com>

ctdb-daemon: Improve readability of code by nesting if-statements

ctdb_sys_have_ip() should only be run if if do_publicipcheck is set.
This is clearer if written as 2 nested if-statements rather than as a
lazy conjuction.

Pair-programmed-with: Amitay Isaacs <amitay@gmail.com>
Signed-off-by: Martin Schwenke <martin@meltin.net>
Reviewed-by: Amitay Isaacs <amitay@gmail.com>

libads: record service ticket endtime for sealed ldap connections

When a ticket is obtained for binding a signed/sealed ldap connection,
its liftime should be recorded in the ads struct, in order to enable
reuse of the connection.

BUG: https://bugzilla.samba.org/show_bug.cgi?id=11267

Signed-off-by: Uri Simchoni <urisimchoni@gmail.com>
Reviewed-by: Jeremy Allison <jra@samba.org>
Reviewed-by: Ralph Böhme <rb@sernet.de>
Autobuild-User(master): Jeremy Allison <jra@samba.org>
Autobuild-Date(master): Wed May 13 04:32:16 CEST 2015 on sn-devel-104

dbwrap: Remove an unused variable

Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>

lib: Remove unused hex_encode()

Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>

ntlm_auth: Remove two uses of hex_encode()

Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>

lib: Make sid_binstring_hex use TALLOC

talloc_tos() is better than plain malloc...

Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>

lib: Simplify sid_binstring_hex()

Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>

Add IPv6 support for determining FQDN during ADS join.

Signed-off-by: David Holder <david.holder@erion.co.uk>
Reviewed-by: Jeremy Allison <jra@samba.org>
Reviewed-by: Ralph Böhme <rb@sernet.de>
Autobuild-User(master): Jeremy Allison <jra@samba.org>
Autobuild-Date(master): Tue May 12 23:35:32 CEST 2015 on sn-devel-104

Add IPv6 support to ADS client side LDAP connects. Corrected format for IPv6 LDAP URI. Signed-off-by: David Holder <david.holder@erion.co.uk>

Signed-off-by: David Holder <david.holder@erion.co.uk>
Reviewed-by: Jeremy Allison <jra@samba.org>
Reviewed-by: Ralph Böhme <rb@sernet.de>

Convert all uint32/16/8 to _t in a couple of include files.

Signed-off-by: Richard Sharpe <rsharpe@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
Autobuild-User(master): Jeremy Allison <jra@samba.org>
Autobuild-Date(master): Tue May 12 04:22:55 CEST 2015 on sn-devel-104

Convert all uses of uint32/16/8 to _t in source3/passdb.

Signed-off-by: Richard Sharpe <rsharpe@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>

Convert all (remaining) uses of uint32/16/8 to _t in source3/printing.

Signed-off-by: Richard Sharpe <rsharpe@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>

Convert all uses of uint32/16/8 to _t in source3/registry.

Signed-off-by: Richard Sharpe <rsharpe@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>

Convert all uses of uint32/16/8 to _t in source3/rpc_server.

Signed-off-by: Richard Sharpe <rsharpe@gmail.com>
Reviewed-by: Jeremy Allison <jra@samba.org>

Convert all uses of uint32/16/8 to _t in source3/rpc_client.

Signed-off-by: Richard Sharpe <rsharpe@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>

Convert all uint32/16/8 to _t in source3/rpcclient.

Signed-off-by: Richard Sharpe <rsharpe@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>

Convert all uses of uint32/16/8 to _t in source3/torture.

Signed-off-by: Richard Sharpe <rsharpe@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>

Change all uses of uint32 to uint32_t in vfs.h. This is part of a general cleanup if people approve. This time around get both vfs.h and vfs.c and all instances of uint8, uint16 and uint32.

Signed-off-by: Richard Sharpe <rsharpe@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>

ctdb-daemon: Remove obsolete IPv4 only controls

Signed-off-by: Amitay Isaacs <amitay@gmail.com>
Reviewed-by: Jeremy Allison <jra@samba.org>

ctdb-daemon: Remove older data structure that supports only IPv4 addresses

Signed-off-by: Amitay Isaacs <amitay@gmail.com>
Reviewed-by: Jeremy Allison <jra@samba.org>

ctdb-daemon: Fix typo in debug message

Signed-off-by: Martin Schwenke <martin@meltin.net>
Reviewed-by: Amitay Isaacs <amitay@gmail.com>
Autobuild-User(master): Amitay Isaacs <amitay@samba.org>
Autobuild-Date(master): Sun May 10 06:10:21 CEST 2015 on sn-devel-104

ctdb-daemon: Initialise eventscript status earlier

Don't initialise it after ctdb_event_script_callback_v() may have
short-circuited.  This can stop ctdb_event_script_args() from ever
terminating.

Signed-off-by: Martin Schwenke <martin@meltin.net>
Reviewed-by: Amitay Isaacs <amitay@gmail.com>

ctdb-daemon: Make ctdb_event_script_args() terminate if no scripts

status.done is never set to true unless event_script_callback() is
invoked.  The short-circuit in ctdb_event_script_callback_v() means
that this doesn't happen.  CTDB can't work very well without 00.ctdb
(for tunable initialisation and the like) but it shouldn't get stuck.

So call the callback when there are no scripts in
event_script_callback().

Signed-off-by: Martin Schwenke <martin@meltin.net>
Reviewed-by: Amitay Isaacs <amitay@gmail.com>

ctdb-daemon: Drop interface monitoring

This is done by 10.interace where the monitor event fails when there
is a missing interface.  The in-daemon interface checking adds no
value.

Signed-off-by: Martin Schwenke <martin@meltin.net>
Reviewed-by: Amitay Isaacs <amitay@gmail.com>

ctdb-common: Reimplement external tracing using ctdb_set_helper()

Signed-off-by: Martin Schwenke <martin@meltin.net>
Reviewed-by: Amitay Isaacs <amitay@gmail.com>

ctdb-scripts: Drop update of public address configuration from config.tdb

This isn't used or documented anywhere.

2 differing points of view:

* This is a very good idea but it should probably be generalised to
  cover more configuration items.  This would end up like the Samba
  registry configuration and would use a tool to support setting
  configuration values.

* If people really want to update configuration while a node is down
  then they should fix the configuration before bringing up that node.

Signed-off-by: Martin Schwenke <martin@meltin.net>
Reviewed-by: Amitay Isaacs <amitay@gmail.com>

ctdb-recoverd: Short circuit takeover run if no nodes are RUNNING

If all nodes are still in, say, FIRST_RECOVERY runstate, then the logs
contain unfortunate noise like:

  recoverd:Failed to find node to cover ip 10.0.2.131

This avoids that by adding an early exit that avoids running
takeover_run_core() when there are no nodes in the
CTDB_RUNSTATE_RUNNING.

To support this add the runstate to the ipflags structure.  There are
clearly other ways of hacking this but this seems the simplest.

Signed-off-by: Martin Schwenke <martin@meltin.net>
Reviewed-by: Amitay Isaacs <amitay@gmail.com>

ctdb-recoverd: Remove redundant condition when checking recovery lock

It isn't possible to hold the recovery lock without having a lock file
set.

This is part of a goal to generalise the recovery lock mechanism to
just use a helper program, which may use a lock file or may use
something else.

Signed-off-by: Martin Schwenke <martin@meltin.net>
Reviewed-by: Amitay Isaacs <amitay@gmail.com>