]>
git.ipfire.org Git - thirdparty/samba.git/log
Ronnie Sahlberg [Sun, 6 May 2007 18:41:12 +0000 (04:41 +1000)]
add a test in the function that checks whether the cluster needs
recovery or not that all active nodes are in normal mode.
If we discover that some node is still in recoverymode it may indicate
that a previous recovery ended prematurely and thus we should start a
new recovery
(This used to be ctdb commit
c15517872e6c98c8c425a8d47d2b348ecb0620b0 )
Ronnie Sahlberg [Sun, 6 May 2007 02:46:56 +0000 (12:46 +1000)]
update a comment to be more desciptive
(This used to be ctdb commit
96082c54d830974bf9a4d5bad33ad60379a85798 )
Ronnie Sahlberg [Sun, 6 May 2007 00:51:25 +0000 (10:51 +1000)]
change a lot of printf into debug statements
(This used to be ctdb commit
6edb9149c7eb36da47e4e6a9dd3ede22263ce3f9 )
Ronnie Sahlberg [Sun, 6 May 2007 00:42:18 +0000 (10:42 +1000)]
break out the code to update all nodes to the new vnnmap into a helper
function
(This used to be ctdb commit
81d39177949b54715710907d14ddc888dc09b064 )
Ronnie Sahlberg [Sun, 6 May 2007 00:38:44 +0000 (10:38 +1000)]
create a helper function for recovery to push all local databases out
onto the remote nodes
(This used to be ctdb commit
1ba76d374652cfa29e56fb77c7190349e42d3bcc )
Ronnie Sahlberg [Sun, 6 May 2007 00:30:18 +0000 (10:30 +1000)]
add an extra blank line
(This used to be ctdb commit
75096dde58df6532abbf5b9ebd771e8810156483 )
Ronnie Sahlberg [Sun, 6 May 2007 00:22:13 +0000 (10:22 +1000)]
break the code that repoints dmaster for all local and remote records
into a separate helper function
(This used to be ctdb commit
d5ab30d0ac21e736eb34eaa19bccfee5f0ce7cfb )
Ronnie Sahlberg [Sun, 6 May 2007 00:16:48 +0000 (10:16 +1000)]
create a helper function for recovery that pulls and merges all remote
databases onto the local node
(This used to be ctdb commit
5cecc47449c369f91e83389a94b987ac32b1e3f4 )
Ronnie Sahlberg [Sun, 6 May 2007 00:12:42 +0000 (10:12 +1000)]
create a helper function to make sure the local node that does recovery
has all the databases that exist on any other remote node
(This used to be ctdb commit
0f436e3d40fea6e6a146019b0c664e80e81e88b4 )
Ronnie Sahlberg [Sun, 6 May 2007 00:04:37 +0000 (10:04 +1000)]
add a helper function to create all missing remote databases detected
during recovery
(This used to be ctdb commit
04758c6f7d8f61260be6d2472380cb7904984427 )
Ronnie Sahlberg [Sat, 5 May 2007 23:53:12 +0000 (09:53 +1000)]
break out the setting/clearing of recovery mode into a dedicated helper
function
(This used to be ctdb commit
dba4e4f8aa4f2fde1e9f8d93bdf3a33f7de8ce18 )
Ronnie Sahlberg [Sat, 5 May 2007 22:05:22 +0000 (08:05 +1000)]
dont allocate arrays where we can just return a single integer
(This used to be ctdb commit
07bc338e490e0f7018808a2450bc54863eb88c94 )
Ronnie Sahlberg [Sat, 5 May 2007 21:52:20 +0000 (07:52 +1000)]
dont use arrays where a uint32_t works just as well
(This used to be ctdb commit
843e974b29c93df891ae7cf13323ee960a334f60 )
Ronnie Sahlberg [Sat, 5 May 2007 21:32:16 +0000 (07:32 +1000)]
add a ifdeffed out block to the call.
we really should kill the event in case the call completed before the
timeout so that we can also make timed_out non-static
(This used to be ctdb commit
f297eed589b1d4e188f77f195683365cf91d0e62 )
Ronnie Sahlberg [Sat, 5 May 2007 21:07:47 +0000 (07:07 +1000)]
hte timed_out variable needs to be static and can not be on the stack
since if the command times out and we return from ctdb_control we may
have events that can trigger later which will overwrite data that is no
longer in our stackframe
(This used to be ctdb commit
93942543092be618c0bd8ef68b470b0789bad7ad )
Ronnie Sahlberg [Sat, 5 May 2007 20:58:01 +0000 (06:58 +1000)]
update to rhe recovery daemon
ctdb_ctrl_ calls are timedout due to nodes arriving or leaving the
cluster it crashes the recovery daemon afterwards with a SEGV but no
useful stack backtrace
(This used to be ctdb commit
cd3abc7349e86555ccd87cd47a1dcc2adad2f46c )
Ronnie Sahlberg [Sat, 5 May 2007 20:06:39 +0000 (06:06 +1000)]
in the recover test
start the daemons with explicit socketnames and explicit ip address/port
remove all --socket= from all ctdb_control calls since they are not
needed anymore
(This used to be ctdb commit
593a959d428f5b4a913117a9b5c8fe65a3eb950e )
Ronnie Sahlberg [Sat, 5 May 2007 19:53:15 +0000 (05:53 +1000)]
add support in catdb to dump the content of a specific nodes tdb instead
of traversing the full cluster.
this makes it easier to debug recovery
update the test script for recovery to reflect the newish signatures to
ctdb_control
the catdb control does still segfault however when there are missing
nodes in the cluster as there are toward the end of the recovery test
(This used to be ctdb commit
8de2a97c14a444f817ceb36461314f10c9601ecc )
Ronnie Sahlberg [Sat, 5 May 2007 18:38:41 +0000 (04:38 +1000)]
merge from tridge
(This used to be ctdb commit
08173e3ab77178b9841db0081a51b93291d9e8dc )
Ronnie Sahlberg [Sat, 5 May 2007 18:31:22 +0000 (04:31 +1000)]
add a control to get the pid of a daemon.
this makes it possible to kill a specific daemon in the recover test
script
(This used to be ctdb commit
2fa394b4c80988cb1a6d04b236ec64cc9d9e8a40 )
Andrew Tridgell [Sat, 5 May 2007 07:46:54 +0000 (17:46 +1000)]
merge relevant lib code from samba4
(This used to be ctdb commit
8076a7c7e12da6d59bae31a2e4a0267d87c7b1b3 )
Andrew Tridgell [Sat, 5 May 2007 07:35:28 +0000 (17:35 +1000)]
merged vnn map broadcast from ronnie
(This used to be ctdb commit
c0fa029435fdaa0be006b28eddb6b31beb2ee605 )
Andrew Tridgell [Sat, 5 May 2007 07:19:59 +0000 (17:19 +1000)]
- take advantage of the new EVENT_FD_AUTOCLOSE flag
- use the tdb_chainlock_mark() call to allow us to guarantee forward progress in the ctdb_lockwait code
(This used to be ctdb commit
e201e98aad0fef6a779a80f3b1ae7792953e2d6b )
Andrew Tridgell [Sat, 5 May 2007 07:18:43 +0000 (17:18 +1000)]
allow the events system to be chosen on the command line
(This used to be ctdb commit
2fe976d7a376a763472cc7952a78b6249ce416c8 )
Andrew Tridgell [Sat, 5 May 2007 07:18:06 +0000 (17:18 +1000)]
use the new lib/events autoconf code
(This used to be ctdb commit
fec779711e8c4d6e047d792aee744e60e5a9f67c )
Andrew Tridgell [Sat, 5 May 2007 07:17:25 +0000 (17:17 +1000)]
- added a EVENT_FD_AUTOCLOSE flag that allows you to tell the event system to close the fd automatically when a fd_event is freed. This prevents races which can lead to epoll missing events
- added autoconf rules for automatically building with epoll support
(This used to be ctdb commit
4d113298b26f7163992f2e47429c953bd4f957c9 )
Andrew Tridgell [Sat, 5 May 2007 07:14:33 +0000 (17:14 +1000)]
added tdb_chainlock_mark() call, which can be used to mark a chain locked without actually locking it. This will be used to guarantee forward progress in the ctdb non-blocking lockwait code
(This used to be ctdb commit
2af98c3418496b39106c7282f550049ec8239657 )
Ronnie Sahlberg [Sat, 5 May 2007 06:51:34 +0000 (16:51 +1000)]
merge from tridge
(This used to be ctdb commit
8648104f8d76d22427c14422b126f7e979cc2d95 )
Andrew Tridgell [Sat, 5 May 2007 04:09:46 +0000 (14:09 +1000)]
show number of connected clients in status output
(This used to be ctdb commit
99765bbe327bfe9c43415f4943281458f25be51b )
Ronnie Sahlberg [Sat, 5 May 2007 03:17:26 +0000 (13:17 +1000)]
split the vnn broadcast address into two
one broadcast address for all nodes
and one broadcast address for all nodes in the current vnnmap
update all useage of the old flag to now only broadcast to the vnnmap
except for tools/ctdb_control where it makes more sense to broadcast to
all nodes
(This used to be ctdb commit
dfb65b88cf67ad9d61268c4b47a6d8ae346f47df )
Ronnie Sahlberg [Sat, 5 May 2007 01:46:44 +0000 (11:46 +1000)]
merge from tridge
(This used to be ctdb commit
ea45ffe3fe479d49b6ed47cb545ee662655f6187 )
Andrew Tridgell [Sat, 5 May 2007 01:03:10 +0000 (11:03 +1000)]
added a dumpmemory control, used to find memory leaks
(This used to be ctdb commit
44fdafaf421e3e906796d529aed2f7c5df201b94 )
Andrew Tridgell [Fri, 4 May 2007 22:33:35 +0000 (08:33 +1000)]
- fixed a crash bug after client disconnect in ctdb_control
- added total memory used to ctdb_control status output
(This used to be ctdb commit
a99ffe4372edc63d83d8c8ebf9a60b3413301f5a )
Andrew Tridgell [Fri, 4 May 2007 22:11:54 +0000 (08:11 +1000)]
- added counters for controls in ctdb_control status
(This used to be ctdb commit
858061372fc9902837a1a5b8bcfc0ada58eec193 )
Andrew Tridgell [Fri, 4 May 2007 21:39:23 +0000 (07:39 +1000)]
merged from ronnie
(This used to be ctdb commit
88f0977f303836b50aa9239a9eb3447646bc1e3f )
Ronnie Sahlberg [Fri, 4 May 2007 15:22:30 +0000 (01:22 +1000)]
merge from tridge
(This used to be ctdb commit
62574808ef4dcb76760f1dd2496fbe8e34197c23 )
Andrew Tridgell [Fri, 4 May 2007 12:18:00 +0000 (22:18 +1000)]
added seqnum propogation code to ctdb
(This used to be ctdb commit
be2572b1b09eaaa1ea6a726d60f16996f9407d13 )
Ronnie Sahlberg [Fri, 4 May 2007 07:05:28 +0000 (17:05 +1000)]
merge from tridge
(This used to be ctdb commit
6c8b90cedc67daa89d54db5268fde18bfc20abaf )
Andrew Tridgell [Fri, 4 May 2007 05:33:28 +0000 (15:33 +1000)]
added a ctdb control for enabling the tdb seqnum
(This used to be ctdb commit
c66920d9fb08a4a33418e2c1dcf1fc320fba3761 )
Andrew Tridgell [Fri, 4 May 2007 05:29:10 +0000 (15:29 +1000)]
added a tdb_enable_seqnum() function
(This used to be ctdb commit
1f89da231c6637e339d5da156d6a48340706fe61 )
Ronnie Sahlberg [Fri, 4 May 2007 05:25:57 +0000 (15:25 +1000)]
remove a exit from the test script
(This used to be ctdb commit
4adb61f8270dbd15732bc458d49a66138dd240cc )
Ronnie Sahlberg [Fri, 4 May 2007 05:21:40 +0000 (15:21 +1000)]
recovery daemon
this program is a client to the local ctdb daemon
every second it pulls all vnnmap and nodemaps from all nodes that are
available and checks if a recovery is required
a recovery is required if :
* all nodes do NOT have an identical vnnmap and generation
* all nodes do NOT have an identical nodemap
* there are active nodes that are NOT in the nodemap
* there are nodes in the nodemap that are NOT active
During recovery, the recovery tool will also make sure that all nodes
know about and have created all databases.
(This used to be ctdb commit
2f2650467bac7e8954de7c17cb34f46b0bdbcd26 )
Andrew Tridgell [Fri, 4 May 2007 03:25:30 +0000 (13:25 +1000)]
make catdb take a dbname instead of an id
(This used to be ctdb commit
365346345c33d2f310bb23d0c6ab5c3ed5e6e938 )
Andrew Tridgell [Fri, 4 May 2007 02:18:39 +0000 (12:18 +1000)]
nicer interface to ctdb traverse
(This used to be ctdb commit
e5ce866dcc5037b5069e42bf1e168b646f007b01 )
Ronnie Sahlberg [Fri, 4 May 2007 01:57:45 +0000 (11:57 +1000)]
also verify that the generation id is the same on all the nodes and if
not, trigger a recovery
(This used to be ctdb commit
46b8a66ee70419c153acf45eeec88c1fc8f230ce )
Ronnie Sahlberg [Fri, 4 May 2007 01:51:07 +0000 (11:51 +1000)]
merge from tridge
(This used to be ctdb commit
5724d41352df252151e54ecdb02baa7f82c3aabf )
Andrew Tridgell [Fri, 4 May 2007 01:41:29 +0000 (11:41 +1000)]
- changed the REQ_REGISTER PDU to be a control
- allow controls to know which client invoked them
- added a client_id to clients, so they can be identified remotely
- added the ability to remove registered srvids
- in the list_keys code, register a temp srvid, then remove it afterwards
(This used to be ctdb commit
29603c51cc6d81362532cd8e50f75c8360c5f5ef )
Andrew Tridgell [Fri, 4 May 2007 00:43:39 +0000 (10:43 +1000)]
merge from ronnie
(This used to be ctdb commit
cf23175526d8fde1ef145605fb545584384b6cec )
Ronnie Sahlberg [Thu, 3 May 2007 23:45:53 +0000 (09:45 +1000)]
update getvnnmap control to take a timeout parameter
dont explicitely free the vnnmap pointer in the getvnnmap control this
is freed by the mem_ctx instead
add code to the recoverd to detect when/if recovery is required
veiry that the number of active nodes, the nodemap and the vnn map is
consistent across the entire cluster and if not trigger a recovery
(which right now just prints "we need to do recovery" to the screen.
(This used to be ctdb commit
2b0a207a3748bdb3394dc9fd0d1c344ee1bb0bb5 )
Ronnie Sahlberg [Thu, 3 May 2007 23:01:01 +0000 (09:01 +1000)]
change the signature for ctdb_ctrl_getnodemap() so that a timeout
parameter is added.
change ctdb_get_connected_nodes in the same way
(This used to be ctdb commit
d85f23bcf4c1230225abb2f4a053c70b68d469aa )
Ronnie Sahlberg [Thu, 3 May 2007 22:32:02 +0000 (08:32 +1000)]
ctdb_control should use the provided timeout and not hardcode to 1.0
seconds
(This used to be ctdb commit
03acb2f450578f6195ab2d0a598f6720e33e7cfb )
Ronnie Sahlberg [Thu, 3 May 2007 22:30:18 +0000 (08:30 +1000)]
start working on a recovery daemon
change ctdb_control so it takes a timeval pointer as argument.
this is the timeout. if the node has not responded within hte timeout
ctdb_control will return an error instead of hanging.
if the timeval pointer is NULL then the call will block indefinitely if
there is no response.
this is used for now in the createdb control but all the helpers
ctdb_ctrl_* should probably be updated to take a timeout parameter as
well.
(This used to be ctdb commit
1fe64b04869b17dbf123851b0fe09df8d28a6211 )
Ronnie Sahlberg [Thu, 3 May 2007 07:16:38 +0000 (17:16 +1000)]
merge from tridge
(This used to be ctdb commit
fb8ac93c7dfc11e774ef1ce05b0d0df1de56a621 )
Andrew Tridgell [Thu, 3 May 2007 07:12:23 +0000 (17:12 +1000)]
first version of traverse is working
(This used to be ctdb commit
ecac90cee389a6fa0e9b1efba521e098a24d323f )
Ronnie Sahlberg [Thu, 3 May 2007 06:18:03 +0000 (16:18 +1000)]
cleanup the control "write record"
(This used to be ctdb commit
4dd5c26a21a5dc2b2f76eb23cfeb4df82ba4e956 )
Ronnie Sahlberg [Thu, 3 May 2007 04:42:53 +0000 (14:42 +1000)]
merge from tridge
(This used to be ctdb commit
abf4c9d0cccd937a7036bd03b1b4c6e635c8eb58 )
Andrew Tridgell [Thu, 3 May 2007 03:53:54 +0000 (13:53 +1000)]
merged from ronnie
(This used to be ctdb commit
57a80110ddfd202f8de37297db76dc43a064e476 )
Andrew Tridgell [Thu, 3 May 2007 03:44:27 +0000 (13:44 +1000)]
don't zero beyond packet header unnecessarily
(This used to be ctdb commit
4cf88ca2ce81db8fe10b0dfedb81d99a2bd93328 )
Ronnie Sahlberg [Thu, 3 May 2007 03:30:38 +0000 (13:30 +1000)]
cleanup getnodemap
(This used to be ctdb commit
3867ccf71a167fb82dbc5a3f03f968a325a0c70b )
Ronnie Sahlberg [Thu, 3 May 2007 03:07:34 +0000 (13:07 +1000)]
fixup getdbmap control so it looks a bit nicer
(This used to be ctdb commit
78a4d61cb78da20af5210488e685c91bc3023e90 )
Andrew Tridgell [Thu, 3 May 2007 02:16:03 +0000 (12:16 +1000)]
first stage of efficient non-blocking ctdb traverse
(This used to be ctdb commit
4c23e6f26bde421bb56b55de9d6cd3e319b2be40 )
Andrew Tridgell [Thu, 3 May 2007 01:13:01 +0000 (11:13 +1000)]
merged cleanup from ronnie
(This used to be ctdb commit
971e7f82f3e1e9bfc8d521080c86431fbb59277d )
Ronnie Sahlberg [Thu, 3 May 2007 01:06:24 +0000 (11:06 +1000)]
break set/get vnn map out from ctdb_control and put it in ctdb_recover.c
for the time being
remove all the [de]marshalling and just pass a structure around instead
(This used to be ctdb commit
b1169555ab7015976c0135ff51121cc238f5887c )
Andrew Tridgell [Wed, 2 May 2007 21:48:08 +0000 (07:48 +1000)]
merge from ronnie
(This used to be ctdb commit
bac8eed2c247a7970836c0e6c2b9e2664922f410 )
Ronnie Sahlberg [Wed, 2 May 2007 21:21:29 +0000 (07:21 +1000)]
merge from tridge
(This used to be ctdb commit
ae72aac605d0c68134d65c213d76496e6293ebc4 )
Andrew Tridgell [Wed, 2 May 2007 21:19:29 +0000 (07:19 +1000)]
merge from ronnie
(This used to be ctdb commit
daba707370b28e200d9e55eb81fc819b62f8fbdd )
Ronnie Sahlberg [Wed, 2 May 2007 19:52:45 +0000 (05:52 +1000)]
merge from tridge
(This used to be ctdb commit
2321a2a31e43208f16d6af15ee55a489670de005 )
Andrew Tridgell [Wed, 2 May 2007 19:51:46 +0000 (05:51 +1000)]
fixed a memory leak in the ctdb_control code
(This used to be ctdb commit
70aa77a66bb5f16c93ecb122b92a6e63f6343ab1 )
Andrew Tridgell [Wed, 2 May 2007 19:21:14 +0000 (05:21 +1000)]
enabled built in popt if system doesn't have it
(This used to be ctdb commit
30c8d69267a78e1d3dea0fcc936ee2939bb7a373 )
Ronnie Sahlberg [Wed, 2 May 2007 12:00:48 +0000 (22:00 +1000)]
merge from tridge
(This used to be ctdb commit
17b73a811009588f836c3f9fd1b775d9d504d30c )
Ronnie Sahlberg [Wed, 2 May 2007 11:31:20 +0000 (21:31 +1000)]
do a real recovery by killing a node and then calling the recover
control
(This used to be ctdb commit
f8d97f8b56f0af83a514ec0364eb915e2d5ea2d7 )
Ronnie Sahlberg [Wed, 2 May 2007 11:14:37 +0000 (21:14 +1000)]
recover.sh test script that build a few database and populates them with
a record then performs a recovery which merges all records onto node 0
(This used to be ctdb commit
54a072deb40d8d4a4a9e35bfde0eec3c867211d0 )
Ronnie Sahlberg [Wed, 2 May 2007 11:00:02 +0000 (21:00 +1000)]
add a recover test change alignment for the pull/push db structures
(This used to be ctdb commit
0eb45623ca103e69765ed577ae02e7f8ca777e37 )
Andrew Tridgell [Wed, 2 May 2007 05:11:11 +0000 (15:11 +1000)]
added a builtin fetch function to support samba3 unlocked fetch
(This used to be ctdb commit
8c57a8355a94a7d714b9bec98533bc40a2bc4684 )
Andrew Tridgell [Wed, 2 May 2007 03:34:55 +0000 (13:34 +1000)]
nicer command parsing in ctdb_control
(This used to be ctdb commit
440077ffabb4ce831004b36ac26bd2f8f9b41499 )
Ronnie Sahlberg [Wed, 2 May 2007 03:31:42 +0000 (13:31 +1000)]
discard REQ/REPLY DMASTER when generation id is wrong or when in
recovery mode
(This used to be ctdb commit
bfd6d430f22e38cff1d64510dc447cfbe1dab35f )
Andrew Tridgell [Wed, 2 May 2007 03:29:03 +0000 (13:29 +1000)]
nicer string handling in usage
(This used to be ctdb commit
8c568ada9b46132ebfa452def4f8ba3f11240532 )
Andrew Tridgell [Wed, 2 May 2007 03:25:45 +0000 (13:25 +1000)]
merged from ronnie
(This used to be ctdb commit
77255f93d432829387dc328b0ae726cc9cba8d68 )
Ronnie Sahlberg [Wed, 2 May 2007 03:14:12 +0000 (13:14 +1000)]
remove sleep from the fetch test
add a test script to start 4 daemons to test recovery
(This used to be ctdb commit
09eab12b00934afa5159fc500faf6b07effb2804 )
Ronnie Sahlberg [Wed, 2 May 2007 02:43:35 +0000 (12:43 +1000)]
add a control to create a database
(This used to be ctdb commit
74e489c6737cc79537c7812ea82daafb1b363ec2 )
Ronnie Sahlberg [Wed, 2 May 2007 01:09:10 +0000 (11:09 +1000)]
remove test code in the fetch test to keep the daemons running forever
(so one has lots of tiume to play with ctdb_control)
(This used to be ctdb commit
8fbaf82a7b3f0e62c887e91731ad8c659122c785 )
Ronnie Sahlberg [Wed, 2 May 2007 01:06:58 +0000 (11:06 +1000)]
change the getnodemap control to a more consistent output for whether a
node is connected or not
(This used to be ctdb commit
65c5fe53937a17e1fa6de5739cbd01b982dc49bb )
Ronnie Sahlberg [Wed, 2 May 2007 01:02:04 +0000 (11:02 +1000)]
update some calls to ctdb_control() that were still using the old
signature (flags field)
update some calls to ctdb_get_config() to use the new name
ctdb_ctrl_get_config()
change #include "talloc/talloc.h" to #include "lib/talloc/talloc.h" in
lib/events/events.h
(This used to be ctdb commit
d2cdd87037b9f0c387228d7d4743da4869929c93 )
Ronnie Sahlberg [Wed, 2 May 2007 00:53:29 +0000 (10:53 +1000)]
merge with tridges tree to resolve all conflicts
(This used to be ctdb commit
0f7c6c580ef0de60af68fd22bce36c0c0b2515b0 )
Ronnie Sahlberg [Wed, 2 May 2007 00:37:43 +0000 (10:37 +1000)]
specify which node to perform recovery to when using the recovery
control
(This used to be ctdb commit
c67f8a1783ce6f5af9940d2e22847ddcd939763d )
Ronnie Sahlberg [Wed, 2 May 2007 00:25:14 +0000 (10:25 +1000)]
fix a bug in pushdb control.
we should only write the pushed record if RSN of the new one if higher
than the rsn of the existing one not the other way around
(This used to be ctdb commit
f535008c3755cf84acff7bc73ea05c0901e940ef )
Ronnie Sahlberg [Wed, 2 May 2007 00:20:34 +0000 (10:20 +1000)]
add an initial recovery control to perform samba3 style recovery
this is not optimized at all and copies/merges all records between
databases instead of only those records for which a certain node is
lmaster. (step 7 should later be enhanced to a, delete the database,
push only those records for which the node is lmaster)
(This used to be ctdb commit
509d2c71169e96a8610f9db91293dc7a73c2cc10 )
Andrew Tridgell [Tue, 1 May 2007 21:37:58 +0000 (07:37 +1000)]
new files for updated events system
(This used to be ctdb commit
15d8308e5a0ce04351d70ac3dd25c7698931ebba )
Andrew Tridgell [Tue, 1 May 2007 21:32:04 +0000 (07:32 +1000)]
merge latest versions of lib/replace, lib/talloc, lib/tdb and lib/events into ctdb bzr tree
(This used to be ctdb commit
eaea8a9fa8d2f5e08f3af619fa1008a663f39053 )
Andrew Tridgell [Tue, 1 May 2007 20:50:12 +0000 (06:50 +1000)]
fixed a lib/events bug found by volker
(This used to be ctdb commit
46cd984500cd77d295f059132b4892dbfbc2bcd3 )
Andrew Tridgell [Tue, 1 May 2007 03:25:02 +0000 (13:25 +1000)]
added a hopcount in ctdb_call
(This used to be ctdb commit
36d838801a2a2008c50322cdbfff65a308b1cd1a )
Andrew Tridgell [Mon, 30 Apr 2007 20:34:55 +0000 (06:34 +1000)]
auto-determine listen address by attempting to bind to each address in the cluster in turn
(This used to be ctdb commit
2fab9f96df2e5b5c51c860fd65caf0e926a63e34 )
Andrew Tridgell [Mon, 30 Apr 2007 13:54:06 +0000 (15:54 +0200)]
added attach command in ctdb_control
(This used to be ctdb commit
172ee33306be2ef5ce17a5b9d7fbcc1f265a1b0b )
Andrew Tridgell [Mon, 30 Apr 2007 13:31:40 +0000 (15:31 +0200)]
changed the way set_call and attach are done so that you can safely
attach to databases after the protocol has started. The daemon
broadcasts information on new databases to the other daemons.
This also eliminates the need for the client to know about the hash
between db name and db_id.
(This used to be ctdb commit
3bad91a9d987d4c09fe3322eac23c2733660ad08 )
Andrew Tridgell [Mon, 30 Apr 2007 09:10:21 +0000 (11:10 +0200)]
don't use stderr here - rely on logging
(This used to be ctdb commit
a2a3b254b22dbaafc8c4934d99cc67d6ae4adbd1 )
Andrew Tridgell [Sun, 29 Apr 2007 20:42:23 +0000 (22:42 +0200)]
saner logfile code
testing of ctdbd
(This used to be ctdb commit
05789da5818f8b20f04779b0df5125914d9047f6 )
Ronnie Sahlberg [Sun, 29 Apr 2007 14:58:27 +0000 (00:58 +1000)]
add push/pull of tdb and a control to copy a tdb from one node to
another node
(This used to be ctdb commit
c313daff4c1362cd08a9f682ce04cab312678038 )
Andrew Tridgell [Sun, 29 Apr 2007 14:19:40 +0000 (16:19 +0200)]
yay! finally fixed the bug that volker, ronnie and I have been chasing
for 2 days.
The main bug was in smbd, but there was a secondary (and more subtle)
bug in ctdb that the bug in smbd exposed. When we get send a dmaster
reply, we have to correctly update the dmaster in the recipient even
if the original requst has timed out, otherwise ctdbd can get into a
loop fighting over who will handle a key.
This patch also cleans up the packet allocation, and makes ctdbd
become a real daemon.
(This used to be ctdb commit
59405e59ef522b97d8e20e4b14310a217141ac7c )
Ronnie Sahlberg [Sun, 29 Apr 2007 13:49:27 +0000 (23:49 +1000)]
merge from tridge
(This used to be ctdb commit
a84e9b47a87fc7d4756b4a179aa2ea0bc7c54c78 )
Ronnie Sahlberg [Sun, 29 Apr 2007 12:51:56 +0000 (22:51 +1000)]
add a new "recovery mode" field to ctdb.
while recovery is in progress the daemon will discard all CTDB_REQ_CALL
and rely on clients retransmitting them
add new controls to get/set the recovery mode
(This used to be ctdb commit
41458a61577885ac49150f830e92e93e634c5411 )