IF YOU WOULD LIKE TO GET AN ACCOUNT, please write an
email to Administrator. User accounts are meant only to access repo
and report issues and/or generate pull requests.
This is a purpose-specific Git hosting for
BaseALT
projects. Thank you for your understanding!
Только зарегистрированные пользователи имеют доступ к сервису!
Для получения аккаунта, обратитесь к администратору.
The caller should not dereference lock_ctx after invoking
process_callbacks(), it might be destroyed already.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=11293
Signed-off-by: Stefan Metzmacher <metze@samba.org>
Reviewed-by: Amitay Isaacs <amitay@gmail.com>
Autobuild-User(master): Amitay Isaacs <amitay@samba.org>
Autobuild-Date(master): Fri Jun 12 15:28:57 CEST 2015 on sn-devel-104
We should not dereference lock_ctx after invoking the callback
in the auto_mark == false case. The callback could have destroyed it.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=11293
Signed-off-by: Stefan Metzmacher <metze@samba.org>
Reviewed-by: Amitay Isaacs <amitay@gmail.com>
Let ctdb_lock_request_destructor() take care of the proper cleanup.
If the request if freed from the callback function, then the lock context
should not be freed. Setting request->lctx to NULL takes care of that
in the destructor.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=11293
Signed-off-by: Amitay Isaacs <amitay@gmail.com>
Reviewed-by: Stefan Metzmacher <metze@samba.org>
There is already code in the destructor to correctly remove it from the
pending or the active queue. This also ensures that when lock context
is in pending queue and if the lock request gets freed, the lock context
is correctly removed from the pending queue.
Thanks to Stefan Metzmacher for noticing this and suggesting the fix.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=11293
Signed-off-by: Amitay Isaacs <amitay@gmail.com>
Reviewed-by: Stefan Metzmacher <metze@samba.org>
The code was added to ctdb_lock_context_destructor() to ensure that
the if a lock_ctx gets freed first, the lock_request does not have a
dangling pointer. However, the reverse is also true. When a lock_request
is freed, then lock_ctx should not have a dangling pointer.
In commit 374cbc7b0f, the code for second
condition was dropped causing a regression.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=11293
Signed-off-by: Stefan Metzmacher <metze@samba.org>
Reviewed-by: Amitay Isaacs <amitay@gmail.com>
If the lock request is freed from within the callback, then setting
lock_ctx->request to NULL in ctdb_lock_context_destructor will end up
corrupting memory. In this case, lock_ctx->request could be reallocated
and pointing to something else. This may cause unexpected abort trying
to dereference a NULL pointer.
So, set lock_ctx->request to NULL before processing callbacks.
This avoids the following valgrind problem.
==3636== Invalid write of size 8
==3636== at 0x151F3D: ctdb_lock_context_destructor (ctdb_lock.c:276)
==3636== by 0x58B3618: _talloc_free_internal (talloc.c:993)
==3636== by 0x58AD692: _talloc_free_children_internal (talloc.c:1472)
==3636== by 0x58AD692: _talloc_free_internal (talloc.c:1019)
==3636== by 0x58AD692: _talloc_free (talloc.c:1594)
==3636== by 0x15292E: ctdb_lock_handler (ctdb_lock.c:471)
==3636== by 0x56A535A: epoll_event_loop (tevent_epoll.c:728)
==3636== by 0x56A535A: epoll_event_loop_once (tevent_epoll.c:926)
==3636== by 0x56A3826: std_event_loop_once (tevent_standard.c:114)
==3636== by 0x569FFFC: _tevent_loop_once (tevent.c:533)
==3636== by 0x56A019A: tevent_common_loop_wait (tevent.c:637)
==3636== by 0x56A37C6: std_event_loop_wait (tevent_standard.c:140)
==3636== by 0x11E03A: ctdb_start_daemon (ctdb_daemon.c:1320)
==3636== by 0x118557: main (ctdbd.c:321)
==3636== Address 0x9c5b660 is 96 bytes inside a block of size 120 free'd
==3636== at 0x4C29D17: free (in
/usr/lib64/valgrind/vgpreload_memcheck-amd64-linux.so)
==3636== by 0x58B32D3: _talloc_free_internal (talloc.c:1063)
==3636== by 0x58B3232: _talloc_free_children_internal (talloc.c:1472)
==3636== by 0x58B3232: _talloc_free_internal (talloc.c:1019)
==3636== by 0x58B3232: _talloc_free_children_internal (talloc.c:1472)
==3636== by 0x58B3232: _talloc_free_internal (talloc.c:1019)
==3636== by 0x58AD692: _talloc_free_children_internal (talloc.c:1472)
==3636== by 0x58AD692: _talloc_free_internal (talloc.c:1019)
==3636== by 0x58AD692: _talloc_free (talloc.c:1594)
==3636== by 0x11EC30: daemon_incoming_packet (ctdb_daemon.c:844)
==3636== by 0x136F4A: lock_fetch_callback (ctdb_ltdb_server.c:268)
==3636== by 0x152489: process_callbacks (ctdb_lock.c:353)
==3636== by 0x152489: ctdb_lock_handler (ctdb_lock.c:468)
==3636== by 0x56A535A: epoll_event_loop (tevent_epoll.c:728)
==3636== by 0x56A535A: epoll_event_loop_once (tevent_epoll.c:926)
==3636== by 0x56A3826: std_event_loop_once (tevent_standard.c:114)
==3636== by 0x569FFFC: _tevent_loop_once (tevent.c:533)
==3636== by 0x56A019A: tevent_common_loop_wait (tevent.c:637)
==3636== by 0x56A37C6: std_event_loop_wait (tevent_standard.c:140)
==3636== by 0x11E03A: ctdb_start_daemon (ctdb_daemon.c:1320)
==3636== by 0x118557: main (ctdbd.c:321)
BUG: https://bugzilla.samba.org/show_bug.cgi?id=11293
Signed-off-by: Stefan Metzmacher <metze@samba.org>
Reviewed-by: Amitay Isaacs <amitay@gmail.com>
This avoids rebuilding ctdb after each commit, instead we're using
SAMBA_VERSION_STRING, which may contain "4.3.0pre1-DEVELOPERBUILD",
instead of "4.3.0pre1.GIT.4ccdd39" for the non-install build.
Signed-off-by: Stefan Metzmacher <metze@samba.org>
Reviewed-by: Michael Adam <obnox@samba.org>
Autobuild-User(master): Michael Adam <obnox@samba.org>
Autobuild-Date(master): Mon Jun 8 15:51:46 CEST 2015 on sn-devel-104
Pair-Programmed-With: Stefan Metzmacher <metze@samba.org>
Signed-off-by: Michael Adam <obnox@samba.org>
Signed-off-by: Stefan Metzmacher <metze@samba.org>
While 'which' is a very common tool, on many distros it is not a requirement
that it be installed. 'type' is a shell built-in specified by the Open Group,
and is found in shells like bash, dash, and ksh across multiple OSes.
Signed-off-by: Jose A. Rivera <jarrpa@samba.org>
Reviewed-by: Michael Adam <obnox@samba.org>
Autobuild-User(master): Michael Adam <obnox@samba.org>
Autobuild-Date(master): Fri Jun 5 20:39:47 CEST 2015 on sn-devel-104
For all the records listed in VACUUM_FETCH, migration requests are sent
immediately without waiting. This means there can only be a single
VACUUM_FETCH processing active at a time.
Signed-off-by: Amitay Isaacs <amitay@gmail.com>
Reviewed-by: Michael Adam <obnox@samba.org>
With the processing of one element factored out,
it is more natural to have the actual loop inside the
handler function. This also makes the talloc/free
bracked around the loop more obvious.
Signed-off-by: Michael Adam <obnox@samba.org>
Reviewed-by: Amitay Isaacs <amitay@gmail.com>
Guenther
Signed-off-by: Günther Deschner <gd@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
Autobuild-User(master): Jeremy Allison <jra@samba.org>
Autobuild-Date(master): Sat May 30 05:03:59 CEST 2015 on sn-devel-104
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
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>
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>
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>
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>
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>
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>
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_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>
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
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>
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>
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>
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>
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>
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>
Election packets from the current node are ignored at the beginning of
the function, so this does not need to be checked.
Signed-off-by: Martin Schwenke <martin@meltin.net>
Reviewed-by: Amitay Isaacs <amitay@gmail.com>
No need to just send it to the recovery master.
This reduces the need for main daemon code to know which node is the
recovery master. The end goal is for the main daemon to not need to
know which node is the recovery master - this information would be
stored in the recovery daemon (and subsequently a separate cluster
management daemon).
Signed-off-by: Martin Schwenke <martin@meltin.net>
Reviewed-by: Amitay Isaacs <amitay@gmail.com>
Databases are only pulled by the recovery master, so it can compare
with current node PNN.
Signed-off-by: Martin Schwenke <martin@meltin.net>
Reviewed-by: Amitay Isaacs <amitay@gmail.com>
Unless this node is the recovery master then this is not needed.
Signed-off-by: Martin Schwenke <martin@meltin.net>
Reviewed-by: Amitay Isaacs <amitay@gmail.com>
It isn't used anywhere else and is always re-initialised to 0.
Signed-off-by: Martin Schwenke <martin@meltin.net>
Reviewed-by: Amitay Isaacs <amitay@gmail.com>
They are initialised and updated but the values are never used.
Signed-off-by: Martin Schwenke <martin@meltin.net>
Reviewed-by: Amitay Isaacs <amitay@gmail.com>
