From 4d4db57dd444ed542691d2df7e10c9c43e3686b3 Mon Sep 17 00:00:00 2001
From: Moe Jette <jette1@llnl.gov>
Date: Tue, 12 Oct 2004 17:34:25 +0000
Subject: [PATCH] Define new switch function to re-initialize the state of a
node.
---
doc/html/switchplugins.html | 13 +++++++++++--
src/common/switch.c | 10 ++++++++++
src/common/switch.h | 5 +++++
src/plugins/switch/elan/switch_elan.c | 5 +++++
src/plugins/switch/none/switch_none.c | 5 +++++
src/slurmd/slurmd.c | 2 ++
6 files changed, 38 insertions(+), 2 deletions(-)
diff --git a/doc/html/switchplugins.html b/doc/html/switchplugins.html
index e55334e83d4..e228e483de8 100644
--- a/doc/html/switchplugins.html
+++ b/doc/html/switchplugins.html
@@ -9,7 +9,7 @@
<meta http-equiv="keywords" content="Simple Linux Utility for Resource Management, SLURM, resource management,
Linux clusters, high-performance computing, Livermore Computing">
<meta name="LLNLRandR" content="UCRL-WEB-204324"
-<meta name="LLNLRandRdate" content="29 September 2004">
+<meta name="LLNLRandRdate" content="12 October 2004">
<meta name="distribution" content="global">
<meta name="description" content="Simple Linux Utility for Resource Management">
<meta name="copyright"
@@ -162,6 +162,15 @@ any other job is in the process of terminating.</p>
is initiated. The slurmctld daemon will also request that slurmd supply current
switch state information on a periodic basis.</p>
+<p class="commandline">int switch_p_clear_node_state(void);</p>
+<p style="margin-left:.2in"><b>Description</b>: Initialize node state.
+If any switch state has previously been established for a job, it will be cleared.
+This will be used to establish a "clean" state for the switch on the node upon
+which it is executed.</p>
+<p style="margin-left:.2in"><b>Returns</b>: SLURM_SUCCESS if successful. On failure,
+the plugin should return SLURM_ERROR and set the errno to an appropriate value
+to indicate the reason for failure.</p>
+
<p class="commandline">int switch_p_alloc_node_info(switch_node_info_t *switch_node);</p>
<p style="margin-left:.2in"><b>Description</b>: Allocate storage for a node's switch
state record. It is recommended that the record contain a magic number for validation
@@ -511,7 +520,7 @@ to maintain data format compatibility across different versions of the plugin.</
<a href="mailto:slurm-dev@lists.llnl.gov">slurm-dev@lists.llnl.gov</a>.</p>
<p><a href="http://www.llnl.gov/"><img align=middle src="lll.gif" width="32" height="32" border="0"></a></p>
<p class="footer">UCRL-WEB-204324<br>
-Last modified 29 September 2004</p></td>
+Last modified 12 October 2004</p></td>
</tr>
</table>
</td>
diff --git a/src/common/switch.c b/src/common/switch.c
index 5754b3863b9..bd22a51441f 100644
--- a/src/common/switch.c
+++ b/src/common/switch.c
@@ -79,6 +79,7 @@ typedef struct slurm_switch_ops {
uint32_t nprocs, uint32_t rank);
char * (*switch_strerror) ( int errnum );
int (*switch_errno) ( void );
+ int (*clear_node) ( void );
int (*alloc_nodeinfo) ( switch_node_info_t *nodeinfo );
int (*build_nodeinfo) ( switch_node_info_t nodeinfo );
int (*pack_nodeinfo) ( switch_node_info_t nodeinfo,
@@ -184,6 +185,7 @@ _slurm_switch_get_ops( slurm_switch_context_t c )
"switch_p_job_attach",
"switch_p_strerror",
"switch_p_get_errno",
+ "switch_p_clear_node_state",
"switch_p_alloc_node_info",
"switch_p_build_node_info",
"switch_p_pack_node_info",
@@ -456,6 +458,14 @@ extern char *switch_strerror(int errnum)
* node switch state monitoring functions
* required for IBM Federation switch
*/
+extern int switch_g_clear_node_state(void)
+{
+ if ( switch_init() < 0 )
+ return SLURM_ERROR;
+
+ return (*(g_context->ops.clear_node))();
+}
+
extern int switch_g_alloc_node_info(switch_node_info_t *switch_node)
{
if ( switch_init() < 0 )
diff --git a/src/common/switch.h b/src/common/switch.h
index 71ffee7cd59..a03cfd1852d 100644
--- a/src/common/switch.h
+++ b/src/common/switch.h
@@ -258,6 +258,11 @@ extern int interconnect_attach(switch_jobinfo_t jobinfo, char ***env,
uint32_t nodeid, uint32_t procid, uint32_t nnodes,
uint32_t nprocs, uint32_t rank);
+/*
+ * Clear switch state on this node
+ */
+extern int switch_g_clear_node_state(void);
+
/*
* Allocate storage for a node's switch state record
*/
diff --git a/src/plugins/switch/elan/switch_elan.c b/src/plugins/switch/elan/switch_elan.c
index 0e7ee998331..89d50564278 100644
--- a/src/plugins/switch/elan/switch_elan.c
+++ b/src/plugins/switch/elan/switch_elan.c
@@ -724,6 +724,11 @@ extern char *switch_p_strerror(int errnum)
* node switch state monitoring functions
* required for IBM Federation switch
*/
+extern int switch_p_clear_node_state(void)
+{
+ return SLURM_SUCCESS;
+}
+
extern int switch_p_alloc_node_info(switch_node_info_t *switch_node)
{
return SLURM_SUCCESS;
diff --git a/src/plugins/switch/none/switch_none.c b/src/plugins/switch/none/switch_none.c
index ec0f47db0d1..9e9787308c4 100644
--- a/src/plugins/switch/none/switch_none.c
+++ b/src/plugins/switch/none/switch_none.c
@@ -224,6 +224,11 @@ extern char *switch_p_strerror(int errnum)
* node switch state monitoring functions
* required for IBM Federation switch
*/
+extern int switch_p_clear_node_state(void)
+{
+ return SLURM_SUCCESS;
+}
+
extern int switch_p_alloc_node_info(switch_node_info_t *switch_node)
{
return SLURM_SUCCESS;
diff --git a/src/slurmd/slurmd.c b/src/slurmd/slurmd.c
index 374262c627d..fb87a59b2c8 100644
--- a/src/slurmd/slurmd.c
+++ b/src/slurmd/slurmd.c
@@ -189,6 +189,8 @@ main (int argc, char *argv[])
if (interconnect_node_init() < 0)
fatal("Unable to initialize interconnect.");
+ if (conf->cleanstart && switch_g_clear_node_state())
+ fatal("Unable to clear interconnect state.");
_create_msg_socket();
--
GitLab