Attachment 145393 Details for Bug 211238 – Newer patch

[patch] Newer patch

clumanager-1.2.34-rebootdebug2.patch (text/plain), 8.89 KB, created by Lon Hohberger on 2007-01-11 21:13:46 UTC

(hide)

Description:

Filename:

MIME Type:

Creator: Lon Hohberger

Created: 2007-01-11 21:13:46 UTC

Size: 8.89 KB

patch

obsolete

>Index: doc/man/cludb.8
>===================================================================
>RCS file: /cvs/devel/clumanager/doc/man/cludb.8,v
>retrieving revision 1.12
>retrieving revision 1.13
>diff -u -r1.12 -r1.13
>--- doc/man/cludb.8	4 May 2006 20:56:25 -0000	1.12
>+++ doc/man/cludb.8	11 Jan 2007 18:57:30 -0000	1.13
>@@ -85,6 +85,18 @@
> .IP "cluster%logfacility"
> This changes the logging facility the cluster uses.  See
> /usr/include/sys/syslog.h for a list of valid facilities.
>+.IP "cluster%reboot_debug"
>+In some cases, clumanager will reboot a node.  Setting this flag to "yes"
>+will cause clumanager to write out /tmp/clumanager-reboot-debug prior
>+to rebooting the machine, which contains the source code file, line, and
>+function in the clumanager source code where the reboot was requested.
>+(yes/no; default=no).  If you think clumanager is rebooting your machine
>+without a reason, set this to "yes" and check for it on the next reboot.
>+You can then either browse the source file, contact Red Hat Support, or
>+post the relevant information to linux-cluster@redhat.com if you require
>+more information.  Note that this does NOT trap any reboot call initiated
>+from outside of clumanager, including but not limited to watchdog-initiated
>+reboots.
> .IP "cluster%disk_error_action"
> Action to take if access to the shared state information (raw devices)
> returns errors (ignore/stop/halt/reboot; default=reboot).  Ignore 
>Index: include/sharedstate.h
>===================================================================
>RCS file: /cvs/devel/clumanager/include/sharedstate.h,v
>retrieving revision 1.12
>retrieving revision 1.13
>diff -u -r1.12 -r1.13
>--- include/sharedstate.h	17 Dec 2003 16:43:56 -0000	1.12
>+++ include/sharedstate.h	11 Jan 2007 18:43:29 -0000	1.13
>@@ -202,4 +202,8 @@
> int shared_storage_deinit(void);
> 
> 
>+void do_reboot(int how, const char *func, const char *file, int line);
>+#define REBOOT(arg) do_reboot(arg, __FUNCTION__, __FILE__, __LINE__)
>+
>+
> #endif /* sharedstate.h */
>Index: src/clulib/shared_services.c
>===================================================================
>RCS file: /cvs/devel/clumanager/src/clulib/shared_services.c,v
>retrieving revision 1.9
>retrieving revision 1.11
>diff -u -r1.9 -r1.11
>--- src/clulib/shared_services.c	16 Dec 2004 19:41:21 -0000	1.9
>+++ src/clulib/shared_services.c	11 Jan 2007 21:04:34 -0000	1.11
>@@ -21,7 +21,7 @@
>  * Shared Service Block read/write functions.
>  */
> /*
>- *  $Id: shared_services.c,v 1.9 2004/12/16 19:41:21 lhh Exp $
>+ *  $Id: shared_services.c,v 1.11 2007/01/11 21:04:34 lhh Exp $
>  *
>  *  author: Tim Burke <tburke at redhat.com>
>  *  description: Interface to Service descriptions.
>@@ -42,6 +42,7 @@
> #include <sys/wait.h>
> #include <sys/time.h>
> #include <sys/reboot.h>
>+#include <sched.h>
> #include <signal.h>
> #include <errno.h>
> #include <clu_lock.h>
>@@ -73,7 +74,7 @@
> static int writeServiceBlock(SharedServiceBlock *svcblk);
> void printServiceBlock(SharedServiceBlock *svcblk);
> 
>-static const char *version __attribute__ ((unused)) = "$Id: shared_services.c,v 1.9 2004/12/16 19:41:21 lhh Exp $";
>+static const char *version __attribute__ ((unused)) = "$Id: shared_services.c,v 1.11 2007/01/11 21:04:34 lhh Exp $";
> /*
>  * .............Configurable Parameters...................
>  *
>@@ -104,6 +105,51 @@
> }
> 
> 
>+void
>+do_reboot(int how, const char *func, const char *file, int line)
>+{
>+	char *val = NULL;
>+	int fd;
>+	char buf[128];
>+	time_t now = time(NULL);
>+	struct sched_param s;
>+
>+	CFG_Get("cluster%reboot_debug", NULL, &val);
>+
>+	if (!val || !strlen(val) || val[0]=='n' ||
>+	    ((val[0] != 'y' && atoi(val)==0))) {
>+		reboot(how);
>+		return;
>+	}
>+	
>+	/* Pre-empt EVERYTHING */
>+	memset(&s,0, sizeof(s));
>+	s.sched_priority = 100;
>+	sched_setscheduler(0, SCHED_RR, &s);
>+
>+	fd = open("/tmp/clumanager-reboot-debug",
>+		  O_RDWR|O_SYNC|O_TRUNC|O_CREAT);
>+	if (fd < 0) {
>+		printf("Failed to create file\n");
>+	} else {
>+		snprintf(buf, sizeof(buf),
>+		 	 "%s-> reboot(%d) @ %s:%d in function %s()\n", 
>+		 	 ctime(&now), how, file, line, func);
>+
>+		write(fd, buf, strlen(buf));
>+		fdatasync(fd);
>+		fsync(fd);
>+		close(fd);
>+		sync(); sync(); sync();
>+	}
>+
>+	reboot(how);
>+	/* notreached */
>+	s.sched_priority = 0;
>+	sched_setscheduler(0, SCHED_OTHER, &s);
>+}
>+
>+
> /**
>  * Take action on read/write error from shared state.
>  */
>@@ -124,7 +170,7 @@
> 		
> 	if (!strcasecmp(val, "halt")) {
> 		clulog(LOG_ALERT, "Shared State Error: HALTING");
>-		reboot(RB_HALT_SYSTEM);
>+		REBOOT(RB_HALT_SYSTEM);
> 		return;
> 	}
> 	
>@@ -137,7 +183,7 @@
> 	
> reboot:	
> 	clulog(LOG_ALERT, "Shared State Error: REBOOTING");
>-	reboot(RB_AUTOBOOT);
>+	REBOOT(RB_AUTOBOOT);
> }
> 
> 
>Index: src/clulib/stonith_fence.c
>===================================================================
>RCS file: /cvs/devel/clumanager/src/clulib/stonith_fence.c,v
>retrieving revision 1.19
>retrieving revision 1.20
>diff -u -r1.19 -r1.20
>--- src/clulib/stonith_fence.c	11 Feb 2005 22:22:56 -0000	1.19
>+++ src/clulib/stonith_fence.c	11 Jan 2007 18:43:29 -0000	1.20
>@@ -343,8 +343,7 @@
> 
> 		if (node == clu_node_id) {
> 			clulog(LOG_ALERT, "STONITH: Shooting self!\n");
>-			sync();
>-			reboot(RB_AUTOBOOT);
>+			REBOOT(RB_AUTOBOOT);
> 		}
> 
> 		/*
>Index: src/daemons/clumembd.c
>===================================================================
>RCS file: /cvs/devel/clumanager/src/daemons/clumembd.c,v
>retrieving revision 1.56
>retrieving revision 1.57
>diff -u -r1.56 -r1.57
>--- src/daemons/clumembd.c	22 Jun 2006 15:03:00 -0000	1.56
>+++ src/daemons/clumembd.c	11 Jan 2007 18:43:29 -0000	1.57
>@@ -20,7 +20,7 @@
>  * Consensus/VF-based membership daemon for Red Hat Cluster Manager.
>  */
> /*
>- * $Id: clumembd.c,v 1.56 2006/06/22 15:03:00 lhh Exp $
>+ * $Id: clumembd.c,v 1.57 2007/01/11 18:43:29 lhh Exp $
>  *
>  * Author: Lon Hohberger <lhh at redhat.com>
>  *	   Brian Stevens <bstevens at redhat.com>
>@@ -50,6 +50,7 @@
> #include <termios.h>
> #include <sys/syslog.h>
> #include <membership.h>
>+#include <sharedstate.h>
> #include <sched.h>
> #include <pthread.h>
> 
>@@ -1736,8 +1737,7 @@
> 		if (!debug && __cmp_tv(&maxtime, &diff) == 1) {
> 			clulog(LOG_EMERG, "Failed to send a heartbeat within "
> 			       "failover time - REBOOTING\n");
>-			sync();
>-			reboot(RB_AUTOBOOT);
>+			REBOOT(RB_AUTOBOOT);
> 		}
> 
> 		/*
>@@ -1993,7 +1993,7 @@
> 					       "a heartbeat within failover "
> 					       "time - REBOOTING\n");
> 					sync();
>-					reboot(RB_AUTOBOOT); /* :( */
>+					REBOOT(RB_AUTOBOOT); /* :( */
> 				}
> 			} else {
> 				failed_sends = 0;
>Index: src/daemons/cluquorumd.c
>===================================================================
>RCS file: /cvs/devel/clumanager/src/daemons/cluquorumd.c,v
>retrieving revision 1.57
>retrieving revision 1.58
>diff -u -r1.57 -r1.58
>--- src/daemons/cluquorumd.c	4 May 2006 20:56:25 -0000	1.57
>+++ src/daemons/cluquorumd.c	11 Jan 2007 18:43:29 -0000	1.58
>@@ -621,7 +621,7 @@
> 		if (!clean_cluster_transition) {
> 			clulog(LOG_EMERG, "Quorum lost! "
> 			       "No STONITH devices - REBOOTING NOW");
>-			reboot(RB_AUTOBOOT);
>+			REBOOT(RB_AUTOBOOT);
> 			exit(1);
> 		}
> 
>@@ -1321,7 +1321,7 @@
> 			clulog(LOG_EMERG, "Lost connection to membership: "
> 			       "REBOOTING NOW\n");
> 			sync(); sync(); sync();
>-			reboot(RB_AUTOBOOT);
>+			REBOOT(RB_AUTOBOOT);
> #endif
> 		}
> 		clist_delete(fd);
>Index: src/daemons/cluquorumd_disk.c
>===================================================================
>RCS file: /cvs/devel/clumanager/src/daemons/cluquorumd_disk.c,v
>retrieving revision 1.16
>retrieving revision 1.17
>diff -u -r1.16 -r1.17
>--- src/daemons/cluquorumd_disk.c	21 Sep 2004 18:50:39 -0000	1.16
>+++ src/daemons/cluquorumd_disk.c	11 Jan 2007 18:43:29 -0000	1.17
>@@ -27,6 +27,7 @@
> #include <xmlwrap.h>
> #include <clulog.h>
> #include <sys/reboot.h>
>+#include <sharedstate.h>
> #include <stonith.h>
> #include <stonithapis.h>
> #include <disk_tie.h>
>@@ -396,7 +397,7 @@
> 		    (status_block.ps_state == NODE_DOWN)) {
> #ifdef OLD_BEHAVIOR
> 			clulog(LOG_EMERG, "Disk-TB: Detected I/O Hang! Rebooting NOW!\n");
>-			reboot(RB_AUTOBOOT);
>+			REBOOT(RB_AUTOBOOT);
> #else
> 			clulog(LOG_WARNING, "Disk-TB: Detected I/O Hang!\n");
> #endif
>Index: src/daemons/clusvcmgrd.c
>===================================================================
>RCS file: /cvs/devel/clumanager/src/daemons/clusvcmgrd.c,v
>retrieving revision 1.64
>retrieving revision 1.65
>diff -u -r1.64 -r1.65
>--- src/daemons/clusvcmgrd.c	22 Jun 2006 15:03:01 -0000	1.64
>+++ src/daemons/clusvcmgrd.c	11 Jan 2007 18:43:29 -0000	1.65
>@@ -25,7 +25,7 @@
>  *
>  */
> 
>-/*static const char *version __attribute__ ((unused)) = "$Revision: 1.64 $";*/
>+/*static const char *version __attribute__ ((unused)) = "$Revision: 1.65 $";*/
> 
> #include <stdio.h>
> #include <stdlib.h>
>@@ -980,7 +980,7 @@
> 		       "Failed to stop service %s uncleanly - REBOOTING\n",
> 		       svcName);
> 		sleep(1);
>-		reboot(RB_AUTOBOOT);
>+		REBOOT(RB_AUTOBOOT);
> 	}
> 	return SUCCESS;
> }
>@@ -2926,7 +2926,7 @@
> 					clulog(LOG_EMERG, "Couldn't reconnect "
> 					       "to the quorum daemon! "
> 					       "REBOOTING");
>-					reboot(RB_AUTOBOOT);					
>+					REBOOT(RB_AUTOBOOT);					
> 				}
> 			}
> 		}

Actions: View | Diff

Attachments on bug 211238: 138735 | 145393