1
1
Граф коммитов

210 Коммитов

Автор SHA1 Сообщение Дата
Ralph Castain
f6ba4f6f30 Per discussion with Jeff, an invalid local rank value should never occur - if it does, it could be indicative of deeper problems in the launch procedure. Thus, rather than allowing the launch to proceed, let's abort.
This commit was SVN r20312.
2009-01-22 00:52:46 +00:00
Ralph Castain
8c87e48721 Fix a user-reported bug whereby the -wdir option would only be applied from the last app_context.
This commit was SVN r20306.
2009-01-21 15:52:12 +00:00
Ralph Castain
5d9de3326c Check for valid local/node ranks before using the returned values
This commit was SVN r20304.
2009-01-21 00:54:50 +00:00
Brian Barrett
d3310a5ad1 fixes to get compiling on Red Storm again
This commit was SVN r20252.
2009-01-12 22:30:00 +00:00
Ralph Castain
2778c13fac Continue to refine the timing instrumentation to identify where launch time is being spent
This commit was SVN r20244.
2009-01-12 19:12:58 +00:00
Ralph Castain
007d68becc Make the data on local children and their jobs available globally on both daemons and the HNP. This simply shifts the data structures from the ODLS base to the orte globals area to support subsequent movement of the daemon collective operations from the odls to the grpcomm framework. As that will be a larger change, it will be implemented on a branch and rolled over separately.
This commit was SVN r20228.
2009-01-08 14:25:56 +00:00
Ralph Castain
1bc125c0a7 CID 1131: cleanup a minor memory leak
This commit was SVN r20200.
2009-01-05 15:05:05 +00:00
Jeff Squyres
6d0d8848ac Fix CID 1129: Remove variable that is set but never used.
This commit was SVN r20194.
2009-01-03 15:39:51 +00:00
Ralph Castain
7787f84540 Per the earlier RFC and some discussion at the Dec ORTE design meeting, add the ompi-top tool and all its supporting infrastructure. This includes a new OPAL pstat framework and data type, currently with rather weak support for Mac OSX and pretty complete support for Linux. The Sun team promised to add Solaris support as well.
Also, per chat with Jeff, modified the Makefile.am's of a few orte tools so that they were consistent in the way we generate the ompi-equivalent cmds.

This commit was SVN r20165.
2008-12-22 20:23:05 +00:00
Shiqing Fan
a5281f0434 - 1/4 commit for Windows Visual Studio and CCP support:
CMakeLists and .windows files.
  In contribs preconfigured and precompiled parts.

This commit was SVN r20108.
2008-12-10 20:59:20 +00:00
Ralph Castain
728a24c8ec After considerable patience and help with debugging/testing from Tim M and Jeff S, return a completed and pretty well tested patch of the IOF to the trunk. This commit includes the previously reverted r20074, r20068, and r20064, as well as changes to fix those commits.
Basically, the remaining problem turned out to be:

1. closing stdout/stderr during orte_finalize of mpirun

2. inadvertently setting up a write event on fd = -1

3. devising a scheme to more accurately track when the stdin write event was active vs closed so it only got released once

This passed prelim MTT testing by Jeff and Tim, but should soak for awhile before migrating to 1.3.

This commit was SVN r20106.

The following SVN revision numbers were found above:
  r20064 --> open-mpi/ompi@a07660aea8
  r20068 --> open-mpi/ompi@ec930d14a9
  r20074 --> open-mpi/ompi@2940309613
2008-12-10 20:40:47 +00:00
Ralph Castain
9d7cb82bba Modify the daemon cmd processor to relay and then process the cmd locally. We couldn't do this before due to the daemon's needing to update contact info prior to doing the relay. However, the new routed system plus the inclusion of the nidmap in the launch message now makes this possible.
It is a small launch performance improvement as now we relay the launch cmd across to the next daemon before taking the time to launch our own local procs. Still, it does allow more parallel operations during the launch procedure.

This commit was SVN r20104.
2008-12-10 19:18:36 +00:00
Ralph Castain
e28210d0dc Revert r20074, r20068, and r20064: remove the IOF proc completion code pending further off-trunk work.
This commit was SVN r20089.

The following SVN revision numbers were found above:
  r20064 --> open-mpi/ompi@a07660aea8
  r20068 --> open-mpi/ompi@ec930d14a9
  r20074 --> open-mpi/ompi@2940309613
2008-12-09 17:11:59 +00:00
Ralph Castain
2940309613 Attempt to solve a race condition showing up in some MTT runs. There were three entry points for proc termination info into the ODLS:
1. a direct callback from waitpid - this set the waitpid_fired flag

2. a notify event callback from the IOF - this set the iof complete flag

3. a message via the daemon cmd processor from the proc "de-registering" the sync, thus indicating it was going through MPI_Finalize.

The problem is that these could overlap, with the first two allowing the orted to declare the proc complete before the daemon had responded to #3.

This change forces all three events to flow through the daemon cmd processor, thus ensuring an ordered handling. I'm not certain this will solve the problem, but will await further MTT reports to see. Unfortunately, the problem doesn't show up on any manual or script-based tests I have been able to run, even when I duplicate the exact cmd that fails under MTT.

This commit was SVN r20074.
2008-12-05 04:20:00 +00:00
Ralph Castain
a07660aea8 Bring over the IOF completion changes. This commit fixes the long-occurring problem whereby application procs could, under some circumstances, lose their final prints to stdout/err. The commit includes:
1. coordination of job completion notification to include a requirement for both waitpid detection AND notification that all iof pipes have been closed by the app

2. change of all IOF read and write events to be non-persistent so they can properly be shutdown and restarted only when required

3. addition of a delay (currently set to 10ms) before restarting the stdin read event. This was required to ensure that the stdout, stderr, and stddiag read events had an opportunity to be serviced in scenarios where large files are attached to stdin.

This commit was SVN r20064.
2008-12-03 17:45:42 +00:00
Ralph Castain
9a57db4a81 To support comm_spawn in fully routed environments, daemons need to know the route to all procs in their job family. They already had this information, but were not retaining it. The infrastructure to do so has existed for some time - just never had the time to complete it.
This commit does that by ensuring that daemons retain knowledge of proc location for all procs in their job family. It required a minor change to the ESS API to allow the daemons to update their pidmaps as data was received. In addition, the routed modules have been updated to take advantage of the newly available info, and the encode/decode pidmap utilities have been updated to communicate the required info in the launch message.

This commit was SVN r20022.
2008-11-18 15:35:50 +00:00
Ralph Castain
68423f7544 Partially restore the iof changes - this repairs the initial observation of inconsistent and incomplete output
This commit was SVN r19999.
2008-11-14 20:36:18 +00:00
Ralph Castain
586334d1c8 Per discussion with Tim Mattox, reset the trunk to pre-19991 level for the iof only. I will shortly add a changeset that will repair the one known error where we were incorrectly closing the stdout/err/diag file descriptors when all we wanted to do was close stdin. I will leave out the changes associated with coordinating proc termination due to race conditions IU encounted during MTT testing. I have been unable to replicate those so far, but we hope to resolve it in the near future.
This commit was SVN r19998.
2008-11-14 20:22:36 +00:00
Ralph Castain
555bbf0c02 Fix the iof race conditions wrt proc termination. This is comprised of two sections:
1. modify the iof to track when a proc actually closes all of its open iof output pipes. When this occurs, notify the odls that the proc's iof is complete. This is done via a zero-time event so that we can step out of the read event before processing the notification.

2. in the odls, modify the waitpid callback so it only flags that it was called. Add a function to receive the iof-complete notification, and a function that checks for both iof complete and waitpid callback before declaring a proc fully terminated. This ensures that we read and deliver -all- of the IO prior to declaring the job complete.

Also modified the odls call to orte_iof.close (and the component's implementation) so it only closes stdin, leaving the other io channels alone. This fixes the other half of the known problem.

This should fix the ticket on this subject, but I'll wait to close it pending further testing in the trunk.

This commit was SVN r19991.
2008-11-12 23:32:01 +00:00
Ralph Castain
5889dcd30b Fix a warning reported by Jeff that actually could cause singleton operations to fail. Ensure that the byte object used to init the job map for singleton's is properly initialized.
This commit was SVN r19957.
2008-11-08 01:09:06 +00:00
Jeff Squyres
f4ba25cf3c Remove linking components against ORTE and OPAL libs. This was
removed from all other components long ago; I'm not sure how these
survived.

This commit was SVN r19956.
2008-11-08 00:56:57 +00:00
George Bosilca
d23fe1bb10 Include Ralph's suggestions, i.e. keep the hnp and orted management in sync.
This commit was SVN r19872.
2008-11-01 00:39:46 +00:00
George Bosilca
9528d33e90 Nothing relevant, few indentations and replace tab by spaces.
This commit was SVN r19870.
2008-10-31 22:24:52 +00:00
George Bosilca
ebe87d1842 Apply some suggestions from Ralph and avoid a pretty nasty race condition on the close of the fd.
The problem was that we close the same fd twice, and that meantime the fd could have been reassigned
to some other file or socket.

This commit was SVN r19869.
2008-10-31 22:23:53 +00:00
Ralph Castain
f54fda489e This is a first step towards supporting fully-routed OOB communications:
1. remove direct routed module (hooray!)

2. add radix tree routed module (binomial remains default)

3. remove duplicate data storage - orteds were storing nidmap and pidmap data in odls, everyone else in ess

4. add ess APIs to update nidmap, add new pidmap - used only by orteds for MPI-2 support

5. modify code to eliminate multiple calls to orte_routed.update_route that recreated info already in ess pidmap. Add ess API to lookup that info instead. Modify routed modules to utilize that capability

6. setup new ability to shutdown orteds without sending back an "ack" message to mpirun - not utilized yet, will require some changes to plm terminate_orteds functions in managed environments (coming soon)

Initial tests indicating that fully routing comm via defined routing trees may not actually have a significant cost for operations like IB QP setup. More tests required to confirm.

This will require an autogen...

This commit was SVN r19866.
2008-10-31 21:10:00 +00:00
George Bosilca
0ce76248e8 Close the file descriptors used to push or pull the data to the children.
Without this patch, doing spawn in a loop ended up by exhausting all
available file descriptors pretty quickly. There were about 5 file
descriptors opened per spawned process. Now the number of file
descriptors managed by the process (orted or HNP)
is a lot smaller.

This commit was SVN r19864.
2008-10-31 18:05:28 +00:00
Ralph Castain
6e5d844c36 Roll in the revamped IOF subsystem. Per the devel mailing list email, this is a complete rewrite of the iof framework designed to simplify the code for maintainability, and to support features we had planned to do, but were too difficult to implement in the old code. Specifically, the new code:
1. completely and cleanly separates responsibilities between the HNP, orted, and tool components.

2. removes all wireup messaging during launch and shutdown.

3. maintains flow control for stdin to avoid large-scale consumption of memory by orteds when large input files are forwarded. This is done using an xon/xoff protocol.

4. enables specification of stdin recipients on the mpirun cmd line. Allowed options include rank, "all", or "none". Default is rank 0.

5. creates a new MPI_Info key "ompi_stdin_target" that supports the above options for child jobs. Default is "none".

6. adds a new tool "orte-iof" that can connect to a running mpirun and display the output. Cmd line options allow selection of any combination of stdout, stderr, and stddiag. Default is stdout.

7. adds a new mpirun and orte-iof cmd line option "tag-output" that will tag each line of output with process name and stream ident. For example, "[1,0]<stdout>this is output"

This is not intended for the 1.3 release as it is a major change requiring considerable soak time.

This commit was SVN r19767.
2008-10-18 00:00:49 +00:00
Ralph Castain
15c47a2473 Revise the daemon collective system to handle comm_spawn patterns that cross into new nodes that are not direct children on the routing tree of the HNP.
Refers to ticket #1548. Although this appears to fix the problem, the ticket will be held open pending further test prior to transition to the 1.3 branch.

This commit was SVN r19674.
2008-10-02 20:08:27 +00:00
Ralph Castain
037231fbcb MOdify the node_rank and local_rank fields to be uint16_t so we can handle more than 256 procs/node. Change the type to a defined one so that any future change can be easily done, if required.
This commit was SVN r19637.
2008-09-25 13:39:08 +00:00
Shiqing Fan
04ee20a880 - Mainly type casts. Microsoft VC++ compiler is too strict.
This commit was SVN r19517.
2008-09-08 15:39:30 +00:00
Ralph Castain
4ef9d15d97 Revamp the opal mca paffinity interface. We ran into a problem when we encountered machines that had "holes" in their physical processor layout - e.g., machines that supported "hotplugging", or that had unpopulated sockets. To solve that problem, we had to clarify at the API level where we were describing physical vs logical processor info, and then translate accordingly in the underlying implementation.
See opal/mca/paffinity/paffinity.h for explanation as to the physical vs logical nature of the params used in the API.

Fixes trac:1435

This commit was SVN r19391.

The following Trac tickets were found above:
  Ticket 1435 --> https://svn.open-mpi.org/trac/ompi/ticket/1435
2008-08-21 19:21:28 +00:00
Ralph Castain
dd16e4e4a6 Update process component of odls.
This commit was SVN r19281.
2008-08-13 20:07:59 +00:00
Ralph Castain
913cf04633 Only co-locate debugger daemon if the orted has local children - prevents mpirun from co-locating a daemon when it has no local procs
This commit was SVN r19280.
2008-08-13 20:06:28 +00:00
Ralph Castain
30f37f762d Enable co-location of debugger daemons during initial launch and when debugging a running job.
Provide support for four MPIR extensions that allow specification of debugger daemon executable, argv for the debugger daemon, whether or not to forward debugger daemon IO, and whether or not debugger daemon will piggy-back on ORTE OOB network. Last is not yet implemented.

No change in behavior or operation occurs unless (a) the debugger specifically utilizes the extensions and, for co-locate while running, the user specifically enables the capability via an MCA param. Two of the MPIR extensions supported here are used in a widely-used debugger for a large-scale installation. The other two extensions are new and being utilized in prototype work by several debuggers for possible future release.

This commit was SVN r19275.
2008-08-13 17:47:24 +00:00
Jeff Squyres
0af7ac53f2 Fixes trac:1392, #1400
* add "register" function to mca_base_component_t
   * converted coll:basic and paffinity:linux and paffinity:solaris to
     use this function
   * we'll convert the rest over time (I'll file a ticket once all
     this is committed)
 * add 32 bytes of "reserved" space to the end of mca_base_component_t
   and mca_base_component_data_2_0_0_t to make future upgrades
   [slightly] easier
   * new mca_base_component_t size: 196 bytes
   * new mca_base_component_data_2_0_0_t size: 36 bytes
 * MCA base version bumped to v2.0
   * '''We now refuse to load components that are not MCA v2.0.x'''
 * all MCA frameworks versions bumped to v2.0
 * be a little more explicit about version numbers in the MCA base
   * add big comment in mca.h about versioning philosophy

This commit was SVN r19073.

The following Trac tickets were found above:
  Ticket 1392 --> https://svn.open-mpi.org/trac/ompi/ticket/1392
2008-07-28 22:40:57 +00:00
Ralph Castain
0735d6f1c2 This commit fixes ticket #1414
Cleanup the logic in the odls for when processes terminate. It turns out that we were only going through the kill_proc logic once instead of looping over all local children when we ordered a daemon to kill its local procs. This went unnoticed for some time as for most systems the local procs were terminated anyway when the daemon terminated due to the parent/child relationship.

Solaris is apparently different - the children are not automatically terminated when the parent dies. As a result, it acts as a detector for this bug.

Mucho thanks to Rolf V. for his help in debugging - and to IM for letting me follow his gdb progress in quasi real-time!

This commit was SVN r19044.
2008-07-26 02:54:43 +00:00
Ralph Castain
fdb2408bf2 Rename the osx paffinity component the "posix" component since it really has nothing osx specific in it - it is just a generic posix call to determine #processors. Set the priority low so that both linux and solaris components override it if they build. It shouldn't build in Windows at all.
Modify the odls to remove a (size_t) typecast in front of the num_processors variable just in case it is returned negative. This usually is accompanied by an opal_error, so this shouldn't make any difference - but it is more technically correct.

This commit was SVN r19008.
2008-07-24 01:54:51 +00:00
Lenny Verkhovsky
b4d54dda57 Fixed possible seqf when using RANKFILE, but not all ranks assigned
Fixed allocation of all ranks when using RANKFILE, but not all ranks assigned
Aborting if using RANKFILE, but np wasn't specified a little earlier
Clean mca_rmaps_rank_file_component.debug

This commit was SVN r19004.
2008-07-23 17:44:02 +00:00
Ralph Castain
e3c3d28bf1 Add some more debugging to tell us how many processors were found when setting sched_yield
This commit was SVN r18999.
2008-07-23 15:28:51 +00:00
Ralph Castain
83e7c19d33 Remove deprecated function - this was incorporated into the paffinity framework a long time ago. Fortunately, nobody was actually using it!
This commit was SVN r18990.
2008-07-23 03:43:31 +00:00
Ralph Castain
6135943382 Update the paffinity call in the ODLS so we retrieve the number of processors on the local node, thus allowing us to correctly set the sched_yield parameter.
This commit was SVN r18946.
2008-07-18 19:19:16 +00:00
Lenny Verkhovsky
a812324963 Fixing "paffinity_base_slot_list" environment
This commit was SVN r18900.
2008-07-14 07:10:50 +00:00
Jeff Squyres
583bf425c0 Fixes trac:1383:
Short version: remove opal_paffinity_alone and restore
mpi_paffinity_alone.  ORTE makes various information available for the
MPI layer to decide what it wants to do in terms of processor
affinity.

Details:

 * remove opal_paffinity_alone MCA param; restore mpi_paffinity_alone
   MCA param
 * move opal_paffinity_slot_list param registration to paffinity base
 * ompi_mpi_init() calls opal_paffinity_base_slot_list_set(); if that
   succeeds use that.  If no slot list was set, see if
   mpi_paffinity_alone was set.  If so, bind this process to its Node
   Local Rank (NLR).  The NLR is the ORTE-maintained slot ID; if you
   COMM_SPAWN to a host in this ORTE universe that already has procs
   on it, the NLR for the new job will start at N (not 0).  So this is
   slightly better than mpi_paffinity_alone in the v1.2 series.
 * If a slot list is specified *and* mpi_paffinity_alone is set, we
   display an error and abort.
 * Remove calls from rmaps/rank_file component to register and lookup
   opal_paffinity mca params. 
 * Remove code in orte/odls that set affinities - instead, have them
   just pass a slot_list if it exists. 
 * Cleanup the orte/odls code that determined
   oversubscribed/want_processor as these were just opposites of each
   other.

This commit was SVN r18874.

The following Trac tickets were found above:
  Ticket 1383 --> https://svn.open-mpi.org/trac/ompi/ticket/1383
2008-07-10 21:12:45 +00:00
Ralph Castain
ba5498cdc6 Repair the MPI-2 dynamic operations. This includes:
1. repair of the linear and direct routed modules

2. repair of the ompi/pubsub/orte module to correctly init routes to the ompi-server, and correctly handle failure to correctly parse the provided ompi-server URI

3. modification of orterun to accept both "file" and "FILE" for designating where the ompi-server URI is to be found - purely a convenience feature

4. resolution of a message ordering problem during the connect/accept handshake that allowed the "send-first" proc to attempt to send to the "recv-first" proc before the HNP had actually updated its routes.

Let this be a further reminder to all - message ordering is NOT guaranteed in the OOB

5. Repair the ompi/dpm/orte module to correctly init routes during connect/accept.

Reminder to all: messages sent to procs in another job family (i.e., started by a different mpirun) are ALWAYS routed through the respective HNPs. As per the comments in orte/routed, this is REQUIRED to maintain connect/accept (where only the root proc on each side is capable of init'ing the routes), allow communication between mpirun's using different routing modules, and to minimize connections on tools such as ompi-server. It is all taken care of "under the covers" by the OOB to ensure that a route back to the sender is maintained, even when the different mpirun's are using different routed modules.

6. corrections in the orte/odls to ensure proper identification of daemons participating in a dynamic launch

7. corrections in build/nidmap to support update of an existing nidmap during dynamic launch

8. corrected implementation of the update_arch function in the ESS, along with consolidation of a number of ESS operations into base functions for easier maintenance. The ability to support info from multiple jobs was added, although we don't currently do so - this will come later to support further fault recovery strategies

9. minor updates to several functions to remove unnecessary and/or no longer used variables and envar's, add some debugging output, etc.

10. addition of a new macro ORTE_PROC_IS_DAEMON that resolves to true if the provided proc is a daemon

There is still more cleanup to be done for efficiency, but this at least works.

Tested on single-node Mac, multi-node SLURM via odin. Tests included connect/accept, publish/lookup/unpublish, comm_spawn, comm_spawn_multiple, and singleton comm_spawn.

Fixes ticket #1256

This commit was SVN r18804.
2008-07-03 17:53:37 +00:00
Ralph Castain
9cebe0ca96 Ckpt the bproc support. All compiles now except for PLM module
This commit was SVN r18744.
2008-06-26 03:48:22 +00:00
Ralph Castain
17fcd72b5d Restore bproc code - if someone wants to maintain it, then more power to them...but it would definitely be easier if the old code is in the trunk. This is all .ompi_ignore'd except for me so I can play with making it compile again in my copious free time.
This commit was SVN r18716.
2008-06-24 01:27:22 +00:00
Ralph Castain
0fa9d88009 Set $PWD for the application proc to match the cwd. If the user specifies a working dir via -wdir, this ensures that the enviro variable matches what they get from getcwd. Note that any subsequent calls to chdir in the user's program will break that equivalence - we can only ensure it starts out matching!
This commit was SVN r18709.
2008-06-23 18:25:41 +00:00
Ralph Castain
3b5e80fa61 Shift responsibility for preconnecting the oob to the orte routed framework, which is the only place that knows what needs to be done. Only the direct module will actually do anything - it uses the same algo as the original preconnect function.
This commit was SVN r18677.
2008-06-19 13:48:26 +00:00
Ralph Castain
0532d799d6 Complete implementation of the --without-rte-support configure option. Working with Brian, this has been tested on RedStorm.
Some minor changes to help facilitate debugger support so that both mpirun and yod can operate with it. Still to be completed.

This commit was SVN r18664.
2008-06-18 03:15:56 +00:00
Ralph Castain
a87aa442e3 Remove last remaining reference to iof_flush - it was #if'd out anyway. The existing flush code appears to have several critical problems. Given the impending rework of the IOF subsystem, there is no point in trying to fix it here.
This commit was SVN r18649.
2008-06-11 16:25:46 +00:00