1
1
Граф коммитов

1825 Коммитов

Автор SHA1 Сообщение Дата
Ralph Castain
9635db373d Ensure that we properly exit if the executable isn't found
This commit was SVN r21570.
2009-07-01 03:16:13 +00:00
Josh Hursey
91c4869cdc Use {{{ORTE_*_VERSION}}} instead of {{{OMPI_*_VERSION}}} in orte
This commit was SVN r21558.
2009-06-29 12:53:23 +00:00
Lenny Verkhovsky
e03807a3d1 small patch to extend current rankfile syntax to be compliant with orte_hosts syntax
making it possible to claim relative hosts from the hostfile/scheduler
by using +n# hostname, where  0 <= # < np
ex:
cat ~/work/svn/hpc/dev/test/Rankfile/rankfile
rank 0=+n0 slot=0
rank 1=+n0 slot=1
rank 2=+n1 slot=2
rank 3=+n1 slot=1

This commit was SVN r21557.
2009-06-28 11:20:56 +00:00
Shiqing Fan
656ec00611 Remove the definition of _WIN32_DCOM. It only enables DCOM when _WIN32_WINNT is less then 0x400 (Win 95, 98), and we are supporting 0x502(Win XP) and above in Open MPI. Thanks George for pointing this out.
This commit was SVN r21555.
2009-06-27 23:36:25 +00:00
Ralph Castain
d0a5468deb Continue gradual cleanup of pointer array addressing.
This commit was SVN r21550.
2009-06-26 22:40:28 +00:00
Ralph Castain
8ccc47b152 Don't wakeup mpirun too early - even though we ordered an abort, we have to wait for the procs/daemons to properly terminate, and give the kill_procs command a chance to get out.
This commit was SVN r21549.
2009-06-26 22:40:00 +00:00
Ralph Castain
0a92fe3739 Pull the HNP node from the right index
This commit was SVN r21547.
2009-06-26 21:43:09 +00:00
Ralph Castain
b96a71b62e Enable restart of individual processes upon command via the errmgr callback function. It needs an external application to drive this capability, so normal operations shouldn't be affected.
Does not support MPI applications. More work coming to update daemon accounting on movement of procs across nodes.

This commit was SVN r21545.
2009-06-26 20:54:58 +00:00
George Bosilca
f06198a999 BAD grpcomm now has the ability to execute the modex offline. The MPI process
prepare the send buffer, and post the collective order to the local daemon. It
then register the callback and return fromthe modex exchange. It will only 
wait for this modex completion when the modex_recv is called. Meanwhile, the
daemon will do the allgather.

This commit was SVN r21543.
2009-06-26 20:32:31 +00:00
George Bosilca
6239e714a6 Extract the modex unpacking in it's own function.
This commit was SVN r21542.
2009-06-26 20:29:54 +00:00
George Bosilca
760df7fcb9 Get rid of the static buffer. Instead use directly the user supplied one.
This commit was SVN r21541.
2009-06-26 19:21:27 +00:00
Shiqing Fan
b0b11e8465 Clean up things properly.
This commit was SVN r21539.
2009-06-26 15:20:19 +00:00
Shiqing Fan
6ba25951b4 Add a separate function for reading remote registry, so that it could be easily reused.
Add two options for plm process module, i.e. remote_env_prefix for getting OMPI prefix on remote computer by reading its user environment variable (OPENMPI_HOME), and remote_reg_prefix is similar, but it reads the registry on the remote computer. Reading remote env prefix has a higher priority than reading reg prefix, so that user can use their own installation of OMPI.

This commit was SVN r21538.
2009-06-26 13:35:45 +00:00
Shiqing Fan
5b4b8f8899 Get rid of a bunch of compiler warnings on Windows.
This commit was SVN r21536.
2009-06-26 07:51:39 +00:00
Lenny Verkhovsky
efa800efea removed orphan files in rankfile mapper
This commit was SVN r21532.
2009-06-25 17:14:10 +00:00
Josh Hursey
26363fcc8b Move the end timer to the end of the launch loop for the HNP. This matches the timing in the SLURM component, and makes the timing work when enabling the 'tree' based rsh launch.
This commit was SVN r21528.
2009-06-25 13:28:34 +00:00
Josh Hursey
8705c7edb1 Fix {{{orte_timing}}} for the RSH component of the PLM.
It never set the daemon launch counter before launching.
In the output for 'total job launch time' make the message match that of the SLURM PLM for easier parsing.

This commit was SVN r21527.
2009-06-25 12:37:58 +00:00
Shiqing Fan
07f7fe1a1b Add two CCP params, for user specifying stdout and stderr to files.
This commit was SVN r21521.
2009-06-25 09:03:50 +00:00
Ralph Castain
00fb79567f Shift some setup items from orterun to the ess/hnp module so that any HNP will perform them.
This commit was SVN r21520.
2009-06-25 03:00:53 +00:00
George Bosilca
fabebd140f If the daemon doesn't contain orted we're adding the prefix_dir and
bin_base twice. Create a temporary full_orted_cmd to cope with this
case.

This commit was SVN r21519.
2009-06-24 23:57:32 +00:00
George Bosilca
32f632ad5a Put back the multi-word command line argument for the RSH PLM.
This commit was SVN r21518.
2009-06-24 23:51:53 +00:00
George Bosilca
c0750343b5 Partially revert 21513. Beware of the exception on the orte_launch_agent which
is treated apart in orte_plm_base_setup_orted_cmd.

This commit was SVN r21517.
2009-06-24 23:48:14 +00:00
Ralph Castain
2e98ba3fd0 Complete implementation of regexp launch with static oob ports. Only enabled for SLURM at this time - migration to Torque coming
This commit was SVN r21516.
2009-06-24 20:31:26 +00:00
George Bosilca
53e76eed75 Missed an include.
This commit was SVN r21515.
2009-06-24 20:19:47 +00:00
George Bosilca
35033aeae9 Allow the PLM RSH to follow the routing table with the spawn, instead of the hard
coded binomial approach. Now, the PLM extract the children information from the
routed component, and the startup follow the routed topology. As a side effect,
we can now launch using linear or radix topologies, in addition to the previously
binomial topology.

This commit was SVN r21514.
2009-06-24 19:56:29 +00:00
George Bosilca
8cb8f28d9d When we get a report from an orted about its state, don't use the sender of
the message to update the structures, but instead use the information from
the URI. The reason is that even the launch report messages can get routed.

Deal with the orted_cmd_line in a single location.

This commit was SVN r21513.
2009-06-24 19:51:52 +00:00
George Bosilca
84a953a2a6 Be less verbose.
This commit was SVN r21512.
2009-06-24 19:49:06 +00:00
Shiqing Fan
6750657a97 Change the way that we detect cluster nodes status and clean up the CCP components.
Normally, any non-windows nodes should return  NodeStatus_Unreachable, but that's not true. We have noticed that Scientific Linux cluster nodes, which are in the same subnet as the Windows nodes, return NodeStatus_PendingApproval value, and the RAS CCP takes them as Windows nodes. So let's just use the "Ready" nodes.

This commit was SVN r21510.
2009-06-24 19:15:15 +00:00
Shiqing Fan
9bdf258d3c Add an initial support for remote process launch using WMI (Windows Management Instrumentation) in the environment where Windows 2003/2008 server and HPC pack are not installed. Users in a domain will be able to use domain computers as a cluster with their domain Credentials, the computers may run Windows XP or higher, e.g. Vista, and/or Windows 7 (later), with correct settings and permissions for WMI and Windows firewall.
This commit was SVN r21508.
2009-06-24 18:59:24 +00:00
George Bosilca
addaf7aaf8 Repair the tree spawn. The problem seems to come from the fact
that now the HNP send the messages using the routed component. In the case
of tree spawn, when a intermediary node spawn a child it doesn't know how
to forward a message to it, so when the node-map message is coming from
the HNP (as there is nothing yet in the contact/routing table) the message
is sent back the way it came. As a result the node-map message keeps jumping
between the HNP and the first level orteds.

The solution is to add a new option to the children orte_parent_uri, which
is only set when the orted is _not_ directly spawned by the HNP. When this
option is present on the argument list, the orted will add the parent to
its routing, and force the parent to update his routes (by sending the URI).
With this approach, the routing tree is build in same time as the processes
are spawned, and all messages from the HNP can be routed to the leaves.

However, this is far from an optimal solution. Right now, this so called tree
spawn, only spawn the children in a tree without doing anything about the
"connect back to the HNP" step. The HNP is flooded with reports from all the
orted. The total number of messages is higher than in the non tree startup
scheme, so we do not expect this approach to be scalable in the current
incarnation. A complete overhaul of the tree startup is required in order
improve the scalability. Stay tuned!

This commit was SVN r21504.
2009-06-23 22:10:25 +00:00
George Bosilca
6a00481285 We know what a daemon is there is no need to dig into the nidmap to find it out.
This commit was SVN r21503.
2009-06-23 20:43:45 +00:00
Ralph Castain
0ba845fed2 Continue development of regular expression support by implementing it for slurm launches. Works for both initial (cmd line and non-cmd line) and comm_spawn launch.
Additional work required to fully enable static port support when using cmd line regular expression launch system.

This commit was SVN r21502.
2009-06-23 20:25:38 +00:00
George Bosilca
bca8015b94 Add the proc_get_daemon capability to the bproc launcher.
This commit was SVN r21501.
2009-06-23 20:21:55 +00:00
George Bosilca
7339530061 Remove the prototype of a non-existant function.
This commit was SVN r21500.
2009-06-23 19:50:23 +00:00
George Bosilca
225f2b01c9 Don't release uninitialized objects.
This commit was SVN r21499.
2009-06-23 19:47:58 +00:00
Ralph Castain
7a802a9d3a Move the plm designation to the argv from the env to support those systems not setup to pass env via rsh.
This commit was SVN r21495.
2009-06-22 18:08:45 +00:00
Ralph Castain
c199dbb241 Revert r21480 - we already did open/select the PLM on non-HNP daemons. This commit broke slave launches on Torque and SLURM as it caused the PLM to be open/selected twice.
The open/select of the PLM is done in orte/mca/ess/base/ess_base_std_orted.c. It only is done when the PLM MCA param is set directing a specific PLM be selected. The function

orte_plm_base_orted_append_basic_args

clears the params passed to the daemon of any PLM selection passed to the HNP. Each PLM then adds a PLM directive if-and-only-if backend PLM support is desired. At present, Torque, SLURM, and rsh all specify this support and direct that the backend orted open the "rsh" PLM.

This commit was SVN r21488.

The following SVN revision numbers were found above:
  r21480 --> open-mpi/ompi@ed585bce8a
2009-06-20 03:58:00 +00:00
Camille Coti
ed585bce8a Initialize the PML if we are a non-HNP daemon.
If we do not initialize the PML, non-HNP daemons will not be able to use its functions. For example, RSH needs it when the tree_spawn mode is 
enabled: daemons call orte_pml.remote_spawn() function to spawn their children in the deployment tree. 

This commit was SVN r21480.
2009-06-19 18:50:06 +00:00
Ralph Castain
771ce035a5 Complete implementation of regular expression generator and parser - now handles leading zero's and suffix in node names.
This commit was SVN r21468.
2009-06-18 04:36:00 +00:00
Jeff Squyres
2a5813ac2d Silence a compiler warning.
This commit was SVN r21459.
2009-06-17 12:26:38 +00:00
Ralph Castain
85e55c5087 Make the IOF macros match for debug vs optimized builds
This commit was SVN r21453.
2009-06-16 22:30:53 +00:00
Rolf vandeVaart
633b996a0f Add sys/wait.h so we can compile on Solaris.
This commit was SVN r21451.
2009-06-16 19:48:43 +00:00
Ralph Castain
e9fc0a74fb Silence compiler warnings
This commit was SVN r21445.
2009-06-16 13:34:31 +00:00
Ralph Castain
74bd80afd9 Do not preload binaries or files if the app isn't being executed on this node
This commit was SVN r21444.
2009-06-16 03:12:30 +00:00
Ralph Castain
d1dd8c2653 Ensure we accurately count the number of new daemons to be launched, especially if we are restarting processes.
Have the resilient mapper also setup for new daemons in case the PLM needs them.

This commit was SVN r21437.
2009-06-15 13:55:01 +00:00
Ralph Castain
c0c56e30c9 Add a missing function to the resilient mapper so it defines daemons in case they are needed
This commit was SVN r21428.
2009-06-12 19:48:13 +00:00
Ralph Castain
170327e575 Reorg the rmaps components to collect shared code for byslot and bynode mapping in the base so we quit duplicating it in every mapper
This commit was SVN r21424.
2009-06-12 17:52:17 +00:00
Ralph Castain
1ee4acb247 Cleanup how we handle pointer arrays in the odls base fns to avoid potential segfaults
This commit was SVN r21423.
2009-06-12 17:51:23 +00:00
Ralph Castain
70b8c89b44 Fix slave spawn, which was hanging because the local daemon never saw the slave job report - it doesn't do it in the normal way, and so the slave launch system itself has to "fake it".
Also complete implementation to printout app_context objects so we see all the fields.

This commit was SVN r21408.
2009-06-10 19:01:08 +00:00
Ralph Castain
f24cefe3d2 Shift the check for adequate file descriptors to after we check if this proc is part of the job to be launched - no point in doing the check more often than absolutely required
This commit was SVN r21406.
2009-06-10 15:18:48 +00:00
Ralph Castain
87d7d693f0 Add a notifier call when the oob retries are exceeded so sys admins are aware of the problem
This commit was SVN r21405.
2009-06-10 15:17:16 +00:00
George Bosilca
77a6f27d44 Update the call to orte_plm_base_create_jobid based on the new interface.
This commit was SVN r21393.
2009-06-08 23:53:53 +00:00
Ralph Castain
86d55d7ebf Fix tight loops over comm_spawn by checking to see if the system has enough child procs and file descriptors available before attempting to launch. If not, introduce a 1sec delay and then test again. This provides a chance for the orted to complete processing of proc terminations from other children, hopefully creating room for the new proc(s).
Update the loop_spawn test to remove a sleep so that it runs at max speed, letting the new code catch when we overrun ourselves and wait for room to be cleared for the next comm_spawn.

This commit was SVN r21390.
2009-06-08 18:28:26 +00:00
Shiqing Fan
5a90b3068e Two type casts.
This commit was SVN r21388.
2009-06-07 12:51:46 +00:00
Ralph Castain
0a67bcb653 Minor cleanups
This commit was SVN r21387.
2009-06-06 15:44:00 +00:00
Ralph Castain
ccf6b2cb8c Implement the ability to register callbacks when specified error states occur
This commit was SVN r21386.
2009-06-06 01:15:31 +00:00
Ralph Castain
0336460b0a Continue implementation of resilient operations by supporting reuse of jobids for restarted procs. Ensure that restarted processes have valid node and local ranks, and that node rank values are passed to direct-launched processes.
This commit was SVN r21385.
2009-06-06 01:08:47 +00:00
Ralph Castain
3815bfbba6 Provide a better error message when the oob cannot send a message after exhausting retries, and then have the proc abort so the job doesn't just hang forever.
Since it could be a daemon that needs to abort, cleanup the abort sequence so the daemon can exit as cleanly as possible.

This commit was SVN r21361.
2009-06-02 23:57:12 +00:00
Ralph Castain
30a357bd8d Provide a "progress meter" for launch that outputs progress as we are launching, especially on large jobs. Also, provide a timeout mechanism so that we cleanly abort if we don't get a response from the next daemon in a specified time.
This commit was SVN r21359.
2009-06-02 23:52:02 +00:00
Ralph Castain
303e3a1d39 Add a resilient mapping capability - currently maps by fault groups (if provided), still need to add the remapping capability for failed procs.
This commit was SVN r21350.
2009-06-02 03:23:20 +00:00
Jeff Squyres
df7c387155 I'm (temporarily?) removing this entry because there's no .window file
in this directory and it's causing "make dist" to break.

Shiqing -- is there a missing file in this directory?  If so, please
add it and restore the EXTRA_DIST line I just removed.  Thanks!

This commit was SVN r21340.
2009-06-01 14:07:08 +00:00
Ralph Castain
137104b786 Initial support for CMs - needs to be pruned as CM support develops
This commit was SVN r21335.
2009-05-30 20:57:23 +00:00
Ralph Castain
4a31c65126 Add initial support for CMs
This commit was SVN r21334.
2009-05-30 20:55:55 +00:00
Ralph Castain
a1005a716f Allow CM's to select the default errmgr component. Add support for error function callbacks
This commit was SVN r21333.
2009-05-30 20:43:42 +00:00
Ralph Castain
a95731fc68 Minor update to let apps set their own component selections if desired, while preserving slave behavior
This commit was SVN r21332.
2009-05-30 20:42:23 +00:00
Ralph Castain
4e0223a638 Add the ability to directly launch procs via rsh/ssh. Collect common functions in plm/base. Create a new global param to set assume_same_shell, alias'd back to plm_rsh_assume_same_shell (not deprecated).
This commit was SVN r21328.
2009-05-30 01:10:25 +00:00
Jeff Squyres
5ea1b776f7 Remove a compiler warning about an empty format string. The proper
way to have no abort message is to pass NULL (the errmanager is smart
enough to handle this case and not emit any extra message).

This commit was SVN r21311.
2009-05-28 13:32:37 +00:00
Ralph Castain
17485ca604 Minor mod per Greg Watson, plus some cleanups to make George smile...or at least grimace a little less! :-)
This commit was SVN r21309.
2009-05-28 00:55:01 +00:00
Ralph Castain
cf3a5c44db Modify the xml output per devel-list discussion with Greg Watson
This commit was SVN r21285.
2009-05-27 00:43:54 +00:00
Jeff Squyres
e6a32f13bb Add missing header file
This commit was SVN r21273.
2009-05-26 20:57:44 +00:00
Iain Bason
e7ff2368d6 This fixes trac:1930.
Emit a more informative error message when the file descriptor limit is
reached during an accept() call.  Also, abort when the accept fails to
avoid an infinite loop.

Emit a more informative error message when the help file can't be opened.

This commit was SVN r21271.

The following Trac tickets were found above:
  Ticket 1930 --> https://svn.open-mpi.org/trac/ompi/ticket/1930
2009-05-26 20:03:21 +00:00
Rolf vandeVaart
db04b6ca71 This change does two things. First, do not emit error
messages when delivering a signal (like STOP or CONT)
to a non-existant process.  This fixes trac:1929.
Also, only print one error message in the other cases.

This commit was SVN r21263.

The following Trac tickets were found above:
  Ticket 1929 --> https://svn.open-mpi.org/trac/ompi/ticket/1929
2009-05-22 14:59:27 +00:00
Ralph Castain
cc7620c210 Fix orte-ps so it properly ignores/reports stale HNPs, but continues to provide output on running ones. Add a timeout on the send side of the comm so we don't hang while trying to send the info request to the non-existent HNP.
This commit was SVN r21257.
2009-05-21 02:42:21 +00:00
Jeff Squyres
154dc846b5 Remove extra ";", which caused a bazillion warnings in MTT.
This commit was SVN r21255.
2009-05-20 13:16:31 +00:00
Rainer Keller
5c80033aa2 - Eliminate icc warning w/ regard to __attribute__((__format__)) on
function pointers... Needed checking in opal_check_attributes.m4

This commit was SVN r21254.
2009-05-20 00:39:22 +00:00
Ralph Castain
f139cfd28a Fully enable the use of static ports to minimize connections on mpirun. When static ports are provided, daemons will automatically use routes defined by the selected routed module to callback to mpirun during startup, thus elimating the dedicated daemon-to-mpirun connection. Therefore, the total number of connections on mpirun will equal the fanout of the routed module (instead of #nodes in job).
Add a new tm ess module that exploits this capability.

Update the various plm modules to enable it - just a minor change reflecting an added param to a plm base function.

Additional fixes included:

1. remove an erroneous cleanup of session directories in the tool finalize procedure - tools don't create session directories to begin with!

2. fix a duplicate free when attempting to execute a non-existent app

3. cleanup an typo in the comm utilities 

4. fix comm_spawn - was perturbed by the changes in pack/unpack of orte_job_t to properly support orte-ps

Been tested on slurm and tm machines, using all tests in orte/test/mpi. May run into issue with command line length on large jobs due to inclusion of node info to support static ports - will fix this next with addition of regexp generator to compress that info.

This commit was SVN r21248.
2009-05-16 04:15:55 +00:00
Rainer Keller
b98a095d22 - Similar to r21229, check for return code from
orte_rml_base_update_contact_info

This commit was SVN r21233.

The following SVN revision numbers were found above:
  r21229 --> open-mpi/ompi@9ad9b20847
2009-05-14 00:36:51 +00:00
Rainer Keller
9b7c4c2354 - Update comment
This commit was SVN r21232.
2009-05-14 00:32:43 +00:00
Rainer Keller
b3daf7184a - Fix Coverity CID 1272:
Need to release src in case of error

 - Fix Coverity CID 1271:
   basename is allocated by opal_basename and later overwritten

This commit was SVN r21230.
2009-05-14 00:28:35 +00:00
Rainer Keller
73fd329cbd - Add the proper __opal_attribute_format__(__printf__...) to
declarations.

This commit was SVN r21226.
2009-05-14 00:10:59 +00:00
Ralph Castain
fd5dd9c4cb Ensure we correctly cycle through map_by_slot when mapping leftover procs in rankfile mapper
This commit was SVN r21219.
2009-05-12 15:41:55 +00:00
Shiqing Fan
9f624cd9a2 A small fix, the right one to use in orte.
This commit was SVN r21213.
2009-05-12 09:53:34 +00:00
Ralph Castain
d396f0a6fc Per the discussion on the devel list, move the binding of processes to processors from MPI_Init to process start. This involves:
1. replacing mpi_paffinity_alone with opal_paffinity_alone - for back-compatibility, I have aliased mpi_paffinity_alone to the new param name. This caus
es a mild abstraction break in the opal/mca/paffinity framework - per the devel discussion...live with it. :-) I also moved the ompi_xxx global variable
 that tracked maffinity setup so it could be properly closed in MPI_Finalize to the opal/mca/maffinity framework to avoid an abstraction break.

2. Added code to the odls/default module to perform paffinity binding and maffinity init between process fork and exec. This has been tested on IU's odi
n cluster and works for both MPI and non-MPI apps.

3. Revise MPI_Init to detect if affinity has already been set, and to attempt to set it if not already done. I have *not* tested this as I haven't yet f
igured out a way to do so - I couldn't get slurm to perform cpu bindings, even though it supposedly does do so.

This has only been lightly tested and would definitely benefit from a wider range of evaluation...

This commit was SVN r21209.
2009-05-12 02:18:35 +00:00
Ralph Castain
fa839f4a30 Fix a bug in the rankfile mapper when nooversubscribe is set
This commit was SVN r21208.
2009-05-11 23:44:59 +00:00
Ralph Castain
388292aed5 Fix my old "friend" singleton comm_spawn
This commit was SVN r21207.
2009-05-11 14:53:02 +00:00
Ralph Castain
c45ff0d59f Take the next step towards fully utilizing static ports for the daemons to eliminate the initial "phone home" to mpirun by modifying the orted termination procedure to eliminate the need for a full barrier-like operation. Instead, we add a "onesided" barrier to the grpcomm framework API that releases the orted once it has completed its own contribution to the barrier - i.e., the orteds now exit as the "ack" message rolls up towards mpirun instead of sending the "ack" directly to mpirun.
This causes the orteds in the routing tree to remain alive until all termination "acks" from orteds below them have passed through. Thus, if we use static ports, we no longer require a direct orted-to-mpirun connection.

Also modify the binomial routed module so it conforms to what all the other routed modules do and have all messages pass along the routing tree instead of short-circuiting between orteds. This further reduces the number of ports being opened on backend nodes.

This commit was SVN r21203.
2009-05-11 14:11:44 +00:00
Ralph Castain
c6f0499720 Some cleanups required from last night's commits to resolve some race conditions and ensure we cleanup properly. Also, remove some debug output that was unintentionally left "on" by default.
This commit was SVN r21202.
2009-05-11 14:03:07 +00:00
Ralph Castain
10a694ea43 The current errmgr.register_callback API takes a jobid as one of its argument. The intent was to have the errmgr check the jobid of the job being reported to it and, if it matches the jobid that was registered, call the specified callback function.
Unfortunately, we assign the jobid during the plm.spawn procedure - which means it happens -after- control of the job has passed out of the range of mpirun (or whatever program is spawning the job), so it is too late for that main program to register a callback function. If the main program registers tha callback -after- we return from plm.spawn, then it (a) cannot get a callback for failed-to-start, and (b) will miss the callback if a proc aborts in the time between job launch and the call to errmgr.register_callback.

This commit fixes the problem by adding callback-related fields to the orte_job_t object. Thus, the main program can specify what job states should initiate a callback, what function is to be called, and what data is to be passed back by simply filling in the orte_job_t fields prior to calling plm.spawn.

Also, fully implement the "copy" function for the orte_job_t object.

NOTE: as a result of this change, the errmgr.register_callback API may no longer be of any value.

This commit was SVN r21200.
2009-05-11 03:38:15 +00:00
Ralph Castain
69cd4e9d8a Continue cleanup of opal_pointer_array references to replace direct addressing with opal_pointer_array_get_item.
Also, be much more careful and complete in recovering resources of terminated jobs.

This commit was SVN r21199.
2009-05-11 03:27:57 +00:00
Josh Hursey
d920a302f3 Some more C/R related commits that have been sitting off-trunk for a while.
* Pass the sequence number of the checkpoint along with reference from the global to the local coordinator.
 * 'orte-restart --apponly' now just generates the app context file, and does not run with it. This provides the user the ability to edit the file before launching. 
 * Add a OPAL_CRS_NONE state
 * Split the INC into three distinct parts.
 * Implement a restart mechanism for the 'none' component. If given a context it simply execvp()'s it.

This commit was SVN r21195.
2009-05-08 20:51:13 +00:00
Josh Hursey
5d0607395d A couple of C/R related commits that have been sitting off-trunk for a while.
* Add 'orte-checkpoint -l' option that lists all checkpoints currently available on the system.
 * Add 'orte-restart -i' which prints information regarding the checkpoint targeted for restart.
 * Add ability to extract the timing metadata.
 * Fix show_help() in the orte-checkpoint and orte-restart tools. They should be using the opal versions instead of the orte versions (otherwise nothing is printed).

This commit was SVN r21194.
2009-05-08 19:41:11 +00:00
Rainer Keller
b0754071b7 - For compilation with BLCR and --with-ft=cr, #include <string.h>
This commit was SVN r21185.
2009-05-07 16:14:59 +00:00
Greg Koenig
60485ff95f This is a very large change to rename several #define values from
OMPI_* to OPAL_*.  This allows opal layer to be used more independent
from the whole of ompi.

NOTE: 9 "svn mv" operations immediately follow this commit.

This commit was SVN r21180.
2009-05-06 20:11:28 +00:00
Shiqing Fan
cd565923d3 Completely remove ltdl support for Windows build.
This commit was SVN r21170.
2009-05-05 18:59:13 +00:00
Josh Hursey
8b8bee04d6 It seems that some of the patches were missed in r21131. :(
This patch contains the following items:
 * Fix the flag passed to open() for the read side of the named pipe between the local and app coordinator. There is a race condition when using O_RDWR on a named pipe (not sure how that bug got in there in the first place).
 * Adjust control in the C/R thread timing
 * Clarify return code in BLCR component
 * Allow the user to adjust the max wait time for the named pipes in the FileM local coordinator by using the MCA parameter "snapc_full_max_wait_time" (Default: 20 seconds)
 * If the application terminates while there are active FileM operations, force mpirun to wait on these operations to complete.
 * Allow the user to set the local copy command (Default: cp) via MCA parameter "filem_rsh_cp"
 * Implement the ability to throttle the number of outgoing connections in FileM. At larger scales this type of explicit throttling helps prevent overwhelming the HNP machine. Default: 10, set via MCA parameter: {{{filem_rsh_max_outgoing}}}

This commit was SVN r21167.

The following SVN revision numbers were found above:
  r21131 --> open-mpi/ompi@0deb009225
2009-05-05 16:45:49 +00:00
Ralph Castain
e615af8b80 Silence coverity...
This commit was SVN r21149.
2009-05-04 22:22:47 +00:00
Ralph Castain
fa531a842d Send all xml output over stdout
This commit was SVN r21147.
2009-05-04 18:51:22 +00:00
Ralph Castain
eac027e6bc In reviewing last night's MTT of trunk, discovered that IU was still testing the old plm slurmd module. This was created solely for debugging proposed changes to the slurm module which have long since been integrated across - and thus, the slurmd module hasn't been maintained for quite some time.
Which explains why the tests using that module all failed...sigh.

My bad for not cleaning it out a long time ago.

This commit was SVN r21145.
2009-05-04 11:25:32 +00:00
Ralph Castain
4be24521aa Modify the orte_process_info structure to handle a broader range of process types by replacing the individual booleans with a 32-bit bitmap. Use a set of #define's to define the individual bits, and a set of matching macros to test for them. Update the orte code base to use the macros instead of the booleans.
Minor mod to the ompi layer to use the new #define's - just one-line name replacements.

This commit was SVN r21144.
2009-05-04 11:07:40 +00:00
Rainer Keller
c32516c9a3 - Include errno.h, to get MTT for sun to run through
This commit was SVN r21143.
2009-05-04 09:13:16 +00:00