Linux Archive - cluster/rgmanager ChangeLog src/resources/clus ...

CVSROOT: /cvs/cluster
Module name: cluster
Changes by: lhh@sourceware.org 2007-11-30 20:06:55

Modified files:
rgmanager : ChangeLog
rgmanager/src/resources: clusterfs.sh fs.sh ocf-shellfuncs
service.sh vm.sh
rgmanager/src/utils: clulog.c clustat.c clusvcadm.c

Log message:
Merges from RHEL5 branch - round 1.

Patches:
http://sourceware.org/cgi-bin/cvsweb.cgi/cluster/rgmanager/ChangeLog.diff?cvsroot=cluster&r1=1.59&r2=1.60
http://sourceware.org/cgi-bin/cvsweb.cgi/cluster/rgmanager/src/resources/clusterfs.sh.diff?cvsroot=cluster&r1=1.19&r2=1.20
http://sourceware.org/cgi-bin/cvsweb.cgi/cluster/rgmanager/src/resources/fs.sh.diff?cvsroot=cluster&r1=1.23&r2=1.24
http://sourceware.org/cgi-bin/cvsweb.cgi/cluster/rgmanager/src/resources/ocf-shellfuncs.diff?cvsroot=cluster&r1=1.4&r2=1.5
http://sourceware.org/cgi-bin/cvsweb.cgi/cluster/rgmanager/src/resources/service.sh.diff?cvsroot=cluster&r1=1.11&r2=1.12
http://sourceware.org/cgi-bin/cvsweb.cgi/cluster/rgmanager/src/resources/vm.sh.diff?cvsroot=cluster&r1=1.7&r2=1.8
http://sourceware.org/cgi-bin/cvsweb.cgi/cluster/rgmanager/src/utils/clulog.c.diff?cvsroot=cluster&r1=1.4&r2=1.5
http://sourceware.org/cgi-bin/cvsweb.cgi/cluster/rgmanager/src/utils/clustat.c.diff?cvsroot=cluster&r1=1.35&r2=1.36
http://sourceware.org/cgi-bin/cvsweb.cgi/cluster/rgmanager/src/utils/clusvcadm.c.diff?cvsroot=cluster&r1=1.22&r2=1.23

--- cluster/rgmanager/ChangeLog 2007/08/30 16:09:38 1.59
+++ cluster/rgmanager/ChangeLog 2007/11/30 20:06:55 1.60
@@ -1,3 +1,7 @@
+2007-11-30 Lon Hohberger <lhh at redhat.com>
+ * src/resources/*: Merge from RHEL5 branch.
+ * src/utils/*: Merge from RHEL5 branch.
+
2007-08-30 Lon Hohberger <lhh at redhat.com>
* src/daemons/restree.c, rg_state.c: Fix tree-restart bug
This is another part of #229650
--- cluster/rgmanager/src/resources/clusterfs.sh 2007/05/21 15:58:04 1.19
+++ cluster/rgmanager/src/resources/clusterfs.sh 2007/11/30 20:06:55 1.20
@@ -133,6 +133,18 @@
<content type="string"/>
</parameter>

+ <parameter name="self_fence">
+ <longdesc lang="en">
+ If set and unmounting the file system fails, the node will
+ immediately reboot. Generally, this is used in conjunction
+ with force-unmount support, but it is not required.
+ </longdesc>
+ <shortdesc lang="en">
+ Seppuku Unmount
+ </shortdesc>
+ <content type="boolean"/>
+ </parameter>
+
<parameter name="fsid">
<longdesc lang="en">
File system ID for NFS exports. This can be overridden
@@ -279,9 +291,6 @@
gfs)
return $OCF_SUCCESS
;;
- ocfs|ocfs2)
- return $OCF_SUCCESS
- ;;
*)
ocf_log err "File system type $OCF_RESKEY_fstype not supported"
return $OCF_ERR_ARGS
@@ -798,6 +807,15 @@
esac
fi

+ if [ -n "$mp" ]; then
+ case ${OCF_RESKEY_self_fence} in
+ $YES_STR) self_fence=$YES ;;
+ 1) self_fence=$YES ;;
+ *) self_fence="" ;;
+ esac
+ fi
+
+
#
# Always do this hackery on clustered file systems.
#
@@ -873,8 +891,13 @@
done # while

if [ -n "$umount_failed" ]; then
- ocf_log err "'umount $dev' failed ($mp), error=$ret_val"
+ ocf_log err "'umount $mp' failed, error=$ret_val"

+ if [ "$self_fence" ]; then
+ ocf_log alert "umount failed - REBOOTING"
+ sync
+ reboot -fn
+ fi
return $FAIL
fi

@@ -884,8 +907,20 @@

case $1 in
start)
- startFilesystem
- exit $?
+ declare tries=0
+ declare rv
+
+ while [ $tries -lt 3 ]; do
+ startFilesystem
+ rv=$?
+ if [ rv -eq 0 ]; then
+ exit 0
+ fi
+
+ ((tries++))
+ sleep 3
+ done
+ exit $rv
;;
stop)
stopFilesystem
--- cluster/rgmanager/src/resources/fs.sh 2007/05/21 15:58:04 1.23
+++ cluster/rgmanager/src/resources/fs.sh 2007/11/30 20:06:55 1.24
@@ -535,8 +535,8 @@
do
#echo "spec=$1 dev=$dev tmp_dev=$tmp_dev"
tmp_dev=$(real_device $tmp_dev)
- tmp_mp=$(trim_trailing_slash $tmp_mp)
- mp=$(trim_trailing_slash $mp)
+ tmp_mp=${tmp_mp/%//} #$(trim_trailing_slash $tmp_mp)
+ mp=${mp/%//} #$(trim_trailing_slash $mp)

if [ -n "$tmp_dev" -a "$tmp_dev" = "$dev" ]; then
#
@@ -1149,6 +1149,9 @@
umount_failed=
done=$YES
;;
+ $FAIL)
+ return $FAIL
+ ;;
$YES)
sync; sync; sync
ocf_log info "unmounting $mp"
--- cluster/rgmanager/src/resources/ocf-shellfuncs 2007/01/15 19:52:05 1.4
+++ cluster/rgmanager/src/resources/ocf-shellfuncs 2007/11/30 20:06:55 1.5
@@ -1,5 +1,5 @@
#
-# $Id: ocf-shellfuncs,v 1.4 2007/01/15 19:52:05 lhh Exp $
+# $Id: ocf-shellfuncs,v 1.5 2007/11/30 20:06:55 lhh Exp $
#
# Common helper functions for the OCF Resource Agents supplied by
# heartbeat.
@@ -174,6 +174,10 @@
esac

pretty_echo $__OCF_PRIO "$__OCF_MSG"
+
+ if [ -z "`which clulog 2> /dev/null`" ]; then
+ return 0
+ fi
clulog -p $__LOG_PID -n $__LOG_NAME
-s $__OCF_PRIO_N "$__OCF_MSG"
}
--- cluster/rgmanager/src/resources/service.sh 2007/07/31 18:00:25 1.11
+++ cluster/rgmanager/src/resources/service.sh 2007/11/30 20:06:55 1.12
@@ -77,7 +77,7 @@
<shortdesc lang="en">
Automatic start after quorum formation
</shortdesc>
- <content type="boolean"/>
+ <content type="boolean" default="1"/>
</parameter>

<parameter name="hardrecovery" reconfig="1">
@@ -92,10 +92,10 @@
<shortdesc lang="en">
Reboot if stop phase fails
</shortdesc>
- <content type="boolean"/>
+ <content type="boolean" default="0"/>
</parameter>

- <parameter name="exclusive">
+ <parameter name="exclusive" reconfig="1">
<longdesc lang="en">
If set, this resource group will only relocate to
nodes which have no other resource groups running in the
@@ -109,7 +109,7 @@
<shortdesc lang="en">
Exclusive resource group
</shortdesc>
- <content type="boolean"/>
+ <content type="boolean" default="0"/>
</parameter>

<parameter name="nfslock">
@@ -125,7 +125,7 @@
<shortdesc lang="en">
Enable NFS lock workarounds
</shortdesc>
- <content type="boolean"/>
+ <content type="boolean" default="0"/>
</parameter>

<parameter name="recovery" reconfig="1">
@@ -141,7 +141,7 @@
<shortdesc lang="en">
Failure recovery policy
</shortdesc>
- <content type="string"/>
+ <content type="string" default="restart"/>
</parameter>

<parameter name="depend">
@@ -154,6 +154,32 @@
</shortdesc>
<content type="string"/>
</parameter>
+
+ <parameter name="max_restarts">
+ <longdesc lang="en">
+ Maximum restarts for this service.
+ </longdesc>
+ <shortdesc lang="en">
+ Maximum restarts for this service.
+ </shortdesc>
+ <content type="string" default="0"/>
+ </parameter>
+
+ <parameter name="restart_expire_time">
+ <longdesc lang="en">
+ Restart expiration time
+ </longdesc>
+ <shortdesc lang="en">
+ Restart expiration time. A restart is forgotten
+ after this time. When combined with the max_restarts
+ option, this lets administrators specify a threshold
+ for when to fail over services. If max_restarts
+ is exceeded in this given expiration time, the service
+ is relocated instead of restarted again.
+ </shortdesc>
+ <content type="string" default="0"/>
+ </parameter>
+
</parameters>

<actions>
@@ -229,8 +255,8 @@
exit 0
;;
reconfig)
- exit 0
- ;;
+ exit 0
+ ;;
*)
exit 0
;;
--- cluster/rgmanager/src/resources/vm.sh 2007/08/02 14:53:38 1.7
+++ cluster/rgmanager/src/resources/vm.sh 2007/11/30 20:06:55 1.8
@@ -77,7 +77,39 @@
<shortdesc lang="en">
Automatic start after quorum formation
</shortdesc>
- <content type="boolean"/>
+ <content type="boolean" default="1"/>
+ </parameter>
+
+ <parameter name="hardrecovery" reconfig="1">
+ <longdesc lang="en">
+ If set to yes, the last owner will reboot if this resource
+ group fails to stop cleanly, thus allowing the resource
+ group to fail over to another node. Use with caution; a
+ badly-behaved resource could cause the entire cluster to
+ reboot. This should never be enabled if the automatic
+ start feature is used.
+ </longdesc>
+ <shortdesc lang="en">
+ Reboot if stop phase fails
+ </shortdesc>
+ <content type="boolean" default="0"/>
+ </parameter>
+
+ <parameter name="exclusive" reconfig="1">
+ <longdesc lang="en">
+ If set, this resource group will only relocate to
+ nodes which have no other resource groups running in the
+ event of a failure. If no empty nodes are available,
+ this resource group will not be restarted after a failure.
+ Additionally, resource groups will not automatically
+ relocate to the node running this resource group. This
+ option can be overridden by manual start and/or relocate
+ operations.
+ </longdesc>
+ <shortdesc lang="en">
+ Exclusive resource group
+ </shortdesc>
+ <content type="boolean" default="0"/>
</parameter>

<parameter name="recovery" reconfig="1">
@@ -177,6 +209,51 @@
<content type="string"/>
</parameter>

+ <parameter name="migrate">
+ <longdesc lang="en">
+ Migration type live or pause, default = live.
+ </longdesc>
+ <shortdesc lang="en">
+ Migration type live or pause, default = live.
+ </shortdesc>
+ <content type="string" default="live"/>
+ </parameter>
+
+ <parameter name="depend">
+ <longdesc lang="en">
+ Top-level service this depends on, in "service:name" format.
+ </longdesc>
+ <shortdesc lang="en">
+ Service dependency; will not start without the specified
+ service running.
+ </shortdesc>
+ <content type="string"/>
+ </parameter>
+
+ <parameter name="max_restarts" reconfig="1">
+ <longdesc lang="en">
+ Maximum restarts for this service.
+ </longdesc>
+ <shortdesc lang="en">
+ Maximum restarts for this service.
+ </shortdesc>
+ <content type="string" default="0"/>
+ </parameter>
+
+ <parameter name="restart_expire_time" reconfig="1">
+ <longdesc lang="en">
+ Restart expiration time
+ </longdesc>
+ <shortdesc lang="en">
+ Restart expiration time. A restart is forgotten
+ after this time. When combined with the max_restarts
+ option, this lets administrators specify a threshold
+ for when to fail over services. If max_restarts
+ is exceeded in this given expiration time, the service
+ is relocated instead of restarted again.
+ </shortdesc>
+ <content type="string" default="0"/>
+ </parameter>

</parameters>

@@ -202,11 +279,11 @@
</actions>

<special tag="rgmanager">
- 
- <attributes maxinstances="1" destroy_on_delete="0" init_on_add="0"/>
+ 
+ <attributes maxinstances="1" destroy_on_delete="0" init_on_add="0"/>
</special>
</resource-agent>
EOT
@@ -220,7 +297,7 @@
# controlled externally; the external monitoring app
# should.
#
- declare cmdline="restart="never""
+ declare cmdline="on_shutdown="destroy" on_reboot="destroy" on_crash="destroy""
declare varp val temp

#
@@ -260,6 +337,8 @@
path)
cmdline="$cmdline --path="$val""
;;
+ migrate)
+ ;;
*)
cmdline="$cmdline $varp="$val""
;;
@@ -285,7 +364,7 @@
#
declare cmdline

- do_status && return 0
+ status && return 0

cmdline="`build_xm_cmdline`"

@@ -314,7 +393,7 @@
while [ $timeout -gt 0 ]; do
sleep 5
((timeout -= 5))
- do_status || return 0
+ status || return 0
while read dom state; do
#
# State is "stopped". Kill it.
@@ -355,45 +434,12 @@
#
do_status()
{
- declare line
-
- xm list $OCF_RESKEY_name &> /dev/null
- if [ $? -eq 0 ]; then
- return $OCF_SUCCESS
- fi
- xm list migrating-$OCF_RESKEY_name &> /dev/null
- if [ $? -eq 1 ]; then
- return $OCF_NOT_RUNNING
- fi
-
- return $OCF_ERR_GENERIC
-
-### NOT REACHED ###
-
- # virsh doesn't handle migrating domains right now
- # When this gets fixed, we need to revisit this status
- # function.
-
- line=$(virsh domstate $OCF_RESKEY_name)
- if [ "$line" = "" ]; then
- return $OCF_NOT_RUNNING
+ xm list $OCF_RESKEY_name &> /dev/null
+ if [ $? -eq 0 ]; then
+ return 0
fi
-
- if [ "$line" = "blocked" ]; then
- return $OCF_SUCCESS
- elif [ "$line" = "running" ]; then
- return $OCF_SUCCESS
- elif [ "$line" = "in shutdown" ]; then
- return $OCF_SUCCESS
- elif [ "$line" = "shut off" ]; then
- return $OCF_NOT_RUNNING
- fi
-
- #
- # Crashed or paused
- #
-
- return $OCF_ERR_GENERIC
+ xm list migrating-$OCF_RESKEY_name &> /dev/null
+ return $?
}

@@ -412,9 +458,13 @@
migrate()
{
declare target=$1
- declare errstr rv
+ declare errstr rv migrate_opt
+
+ if [ "$OCF_RESKEY_migrate" = "live" ]; then
+ migrate_opt="-l"
+ fi

- err=$(xm migrate $OCF_RESKEY_name $target 2>&1 | head -1)
+ err=$(xm migrate $migrate_opt $OCF_RESKEY_name $target 2>&1 | head -1)
rv=$?

if [ $rv -ne 0 ]; then
@@ -446,7 +496,7 @@
exit $?
;;
kill)
- do_stop destroy
+ stop destroy
exit $?
;;
recover|restart)
@@ -457,7 +507,7 @@
exit $?
;;
migrate)
- do_migrate $2 # Send VM to this node
+ migrate $2 # Send VM to this node
exit $?
;;
reload)
@@ -478,6 +528,6 @@
;;
*)
echo "usage: $0 {start|stop|restart|status|reload|reconfig|meta-data|validate-all}"
- exit $OCF_ERR_UNIMPLEMENTED
+ exit 1
;;
esac
--- cluster/rgmanager/src/utils/clulog.c 2007/03/27 19:33:20 1.4
+++ cluster/rgmanager/src/utils/clulog.c 2007/11/30 20:06:55 1.5
@@ -92,9 +92,6 @@
if (argc < 4)
usage(argv[0]);

- /* ../daemons/log.c */
- configure_logging(-1);
-
while ((opt = getopt(argc, argv, "f:l:s:hp:n:")) != -1) {
switch (opt) {
case 'l':
@@ -133,9 +130,10 @@
if (!cmdline_loglevel) {
/*
* Let's see what loglevel the SM is running at.
- * TODO Get rgmgr log level
+ * If ccsd's not available, use default.
*/
- clu_set_loglevel(LOGLEVEL_DFLT);
+ if (configure_logging(-1) < 0)
+ clu_set_loglevel(LOGLEVEL_DFLT);
}
result = clulog_pid(severity, pid, progname, logmsg);
free(progname);
--- cluster/rgmanager/src/utils/clustat.c 2007/09/19 10:56:08 1.35
+++ cluster/rgmanager/src/utils/clustat.c 2007/11/30 20:06:55 1.36
@@ -21,6 +21,7 @@
#define FLAG_RGMGR 0x4
#define FLAG_NOCFG 0x8 /* Shouldn't happen */
#define FLAG_QDISK 0x10
+#define FLAG_RGMAST 0x20 /* for RIND */

#define RG_VERBOSE 0x1

@@ -30,6 +31,8 @@

int running = 1;
+int dimx = 80, dimy = 24, stdout_is_tty = 0;
+int rgmanager_master_present = 0;

void
term_handler(int sig)
@@ -44,6 +47,28 @@
} rg_state_list_t;

+int
+rg_name_sort(const void *left, const void *right)
+{
+ return strcmp(((rg_state_t *)left)->rs_name,
+ ((rg_state_t *)right)->rs_name);
+}
+
+
+int
+member_id_sort(const void *left, const void *right)
+{
+ cman_node_t *l = (cman_node_t *)left;
+ cman_node_t *r = (cman_node_t *)right;
+
+ if (l->cn_nodeid < r->cn_nodeid)
+ return -1;
+ if (l->cn_nodeid > r->cn_nodeid)
+ return 1;
+ return 0;
+}
+
+
void
flag_rgmanager_nodes(cluster_member_list_t *cml)
{
@@ -55,7 +80,7 @@
struct timeval tv;

if (msg_open(MSG_SOCKET, 0, 0, &ctx, 10) < 0) {
- perror("msg_open");
+ //perror("msg_open");
return;
}

@@ -121,6 +146,10 @@
if (cml->cml_members[n].cn_nodeid != msgp->gh_arg1)
continue;
cml->cml_members[n].cn_member |= FLAG_RGMGR;
+ if (msgp->gh_arg2) {
+ rgmanager_master_present = 1;
+ cml->cml_members[n].cn_member |= FLAG_RGMAST;
+ }
}

free(msgp);
@@ -147,7 +176,7 @@
struct timeval tv;

if (msg_open(MSG_SOCKET, 0, 0, &ctx, 10) < 0) {
- perror("msg_open");
+ //perror("msg_open");
return NULL;
}

@@ -248,6 +277,9 @@
return NULL;
}

+ qsort(rsl->rgl_states, rsl->rgl_count, sizeof(rg_state_t),
+ rg_name_sort);
+
return rsl;
}

@@ -270,9 +302,11 @@
sleep(1);

x = 0;
+ memset(buf, 0, sizeof(buf));
+
while (++x) {
name = NULL;
- snprintf(buf, sizeof(buf),
+ snprintf(buf, sizeof(buf)-1,
"/cluster/clusternodes/clusternode[%d]/@name", x);

if (ccs_get(desc, buf, &name) != 0)
@@ -307,7 +341,7 @@
free(name);

/* Add node ID */
- snprintf(buf, sizeof(buf),
+ snprintf(buf, sizeof(buf)-1,
"/cluster/clusternodes/clusternode[%d]/@nodeid", x);
if (ccs_get(desc, buf, &name) == 0) {
nodes[x-1].cn_nodeid = atoi(name);
@@ -320,6 +354,9 @@
ccs_disconnect(desc);
ret->cml_members = nodes;

+ qsort(ret->cml_members, ret->cml_count, sizeof(cman_node_t),
+ member_id_sort);
+
return ret;
}

@@ -413,68 +450,71 @@

void
-_txt_rg_state(rg_state_t *rs, cluster_member_list_t *members, int flags)
+_txt_rg_state(rg_state_t *rs, cluster_member_list_t *members, int flags,
+ char *fmt_buf, int ns)
{
- char owner[31];
- char flags_string[255] = "";
+ char owner[MAXHOSTNAMELEN+1];
+ char owner_fmt[16];
+ char *name = rs->rs_name, *ptr;
+ int l;
+
+ if (stdout_is_tty) {
+ ptr = strchr(rs->rs_name, ':');
+ if (ptr) {
+ l = (int)(ptr - rs->rs_name);
+ if ((l == 7) && /* strlen("service") == 7 */
+ (strncmp(rs->rs_name, "service", l) == 0))
+ name = ptr+1;
+ }
+ }
+
+ memset(owner, 0, sizeof(owner));
+ memset(owner_fmt, 0, sizeof(owner_fmt));

if (rs->rs_state == RG_STATE_STOPPED ||
rs->rs_state == RG_STATE_DISABLED ||
rs->rs_state == RG_STATE_ERROR ||
rs->rs_state == RG_STATE_FAILED) {

- snprintf(owner, sizeof(owner), "(%-.28s)",
+ snprintf(owner_fmt, sizeof(owner_fmt)-1, "(%%-.%ds)", ns-2);
+ snprintf(owner, sizeof(owner)-1, owner_fmt,
my_memb_id_to_name(members, rs->rs_last_owner));
} else {

- snprintf(owner, sizeof(owner), "%-.30s",
+ snprintf(owner_fmt, sizeof(owner_fmt)-1, "%%-.%ds", ns);
+ snprintf(owner, sizeof(owner)-1, owner_fmt,
my_memb_id_to_name(members, rs->rs_owner));
}
- rg_flags_str(flags_string, sizeof(flags_string), rs->rs_flags, ", ");
- printf(" %-20.20s %-30.30s %-16.16s ",
- rs->rs_name,
+
+ printf(fmt_buf,
+ name,
owner,
rg_state_str(rs->rs_state));
- if(strlen(flags_string))
- printf ("%-30.30s
", flags_string);
- else
- printf("
");
}

void
_txt_rg_state_v(rg_state_t *rs, cluster_member_list_t *members, int flags)
{
- char flags_string[255] = "";
- time_t transtime = rs->rs_transition;
-
- rg_flags_str(flags_string, sizeof(flags_string), rs->rs_flags, ", ");
-
printf("Service Name : %s
", rs->rs_name);
printf(" Current State : %s (%d)
",
rg_state_str(rs->rs_state), rs->rs_state);
- if (rs->rs_flags)
- printf(" Flags : %s (%d)
",
- flags_string, rs->rs_flags);
- else
- printf(" Flags : none (%d)
",
- rs->rs_flags);
printf(" Owner : %s
",
my_memb_id_to_name(members, rs->rs_owner));
printf(" Last Owner : %s
",
my_memb_id_to_name(members, rs->rs_last_owner));
printf(" Last Transition : %s
",
- ctime(&transtime));
+ ctime((time_t *)(&rs->rs_transition)));
}

void
-txt_rg_state(rg_state_t *rs, cluster_member_list_t *members, int flags)
+txt_rg_state(rg_state_t *rs, cluster_member_list_t *members, int flags, char *fmt_buf, int ns)
{
if (flags & RG_VERBOSE)
_txt_rg_state_v(rs, members, flags);
else
- _txt_rg_state(rs, members, flags);
+ _txt_rg_state(rs, members, flags, fmt_buf, ns);
}

@@ -482,12 +522,10 @@
xml_rg_state(rg_state_t *rs, cluster_member_list_t *members, int flags)
{
char time_str[32];
- char flags_string[255] = "";
- time_t transtime = rs->rs_transition;
int x;

/* Chop off newlines */
- ctime_r((time_t *)&transtime, time_str);
+ ctime_r((time_t *)&rs->rs_transition, time_str);
for (x = 0; time_str[x]; x++) {
if (time_str[x] < 32) {
time_str[x] = 0;
@@ -495,15 +533,12 @@
}
}

- printf(" <group name="%s" state="%d" state_str="%s""
- " flags="%d" flags_str="%s""
+ printf(" <group name="%s" state="%d" state_str="%s" "
" owner="%s" last_owner="%s" restarts="%d""
" last_transition="%llu" last_transition_str="%s"/>
",
rs->rs_name,
rs->rs_state,
rg_state_str(rs->rs_state),
- rs->rs_flags,
- rg_flags_str(flags_string, sizeof(flags_string), rs->rs_flags, " "),
my_memb_id_to_name(members, rs->rs_owner),
my_memb_id_to_name(members, rs->rs_last_owner),
rs->rs_restarts,
@@ -512,11 +547,37 @@
}

+void
+build_service_format(char *buf, int buflen, int cols, int *ns)
+{
+ /* Based on 80 columns */
+ int svcsize = 30;
+ int nodesize = 30;
+ int statsize = 14; /* uninitialized */
+ int pad = 6; /* Spaces and such; newline */
+
+ svcsize = (cols - (statsize + pad)) / 2;
+ nodesize = (cols - (statsize + pad)) / 2;
+ if (svcsize > MAXHOSTNAMELEN)
+ svcsize = MAXHOSTNAMELEN;
+ if (nodesize > MAXHOSTNAMELEN)
+ nodesize = MAXHOSTNAMELEN;
+
+ memset(buf, 0, buflen);
+ snprintf(buf, buflen-1, " %%-%d.%ds %%-%d.%ds %%-%d.%ds
",
+ svcsize, svcsize, nodesize, nodesize, statsize,
+ statsize);
+
+ *ns = nodesize;
+}
+
+
int
txt_rg_states(rg_state_list_t *rgl, cluster_member_list_t *members,
char *svcname, int flags)
{
- int x, ret = 0;
+ int x, ret = 0, ns;
+ char fmt_buf[80];

if (!rgl || !members)
return -1;
@@ -524,11 +585,14 @@
if (svcname)
ret = -1;

+ build_service_format(fmt_buf, sizeof(fmt_buf), dimx, &ns);
+
if (!(flags & RG_VERBOSE)) {
- printf(" %-20.20s %-30.30s %-16.16s %-30.30s
",
- "Service Name", "Owner (Last)", "State", "Flags");
- printf(" %-20.20s %-30.30s %-16.16s %-30.30s
",
- "------- ----", "----- ------", "-----", "-----");
+
+ printf(fmt_buf,
+ "Service Name", "Owner (Last)", "State");
+ printf(fmt_buf,
+ "------- ----", "----- ------", "-----");
} else {
printf("Service Information
"
"------- -----------

");
@@ -538,7 +602,7 @@
if (svcname &&
strcmp(rgl->rgl_states[x].rs_name, svcname))
continue;
- txt_rg_state(&rgl->rgl_states[x], members, flags);
+ txt_rg_state(&rgl->rgl_states[x], members, flags, fmt_buf, ns);
if (svcname) {
switch (rgl->rgl_states[x].rs_state) {
case RG_STATE_STARTING:
@@ -551,7 +615,7 @@
}
}
}
-
+
return ret;
}

@@ -604,6 +668,24 @@

void
+txt_cluster_info(cman_cluster_t *ci)
+{
+ time_t now = time(NULL);
+
+ printf("Cluster Status for %s @ %s",
+ ci->ci_name, ctime(&now));
+}
+
+
+void
+xml_cluster_info(cman_cluster_t *ci)
+{
+ printf(" <cluster name="%s" id="%d" generation="%d"/>
",
+ ci->ci_name, ci->ci_number, ci->ci_generation);
+}
+
+
+void
xml_quorum_state(int qs)
{
/* XXX output groupmember attr (carry over from RHCS4) */
@@ -619,15 +701,31 @@
} else {
printf(" groupmember="0"");
}
+
printf("/>
");
}

+void
+build_member_format(char *buf, int buflen, int cols)
+{
+ /* Based on 80 columns */
+ int nodesize = 40;
+
+ nodesize = (cols / 2);
+ if (nodesize > MAXHOSTNAMELEN)
+ nodesize = MAXHOSTNAMELEN;
+
+ memset(buf, 0, buflen);
+ snprintf(buf, buflen-1, " %%-%d.%ds ",
+ nodesize, nodesize);
+}
+

void
-txt_member_state(cman_node_t *node)
+txt_member_state(cman_node_t *node, char *fmt_buf)
{
- printf(" %-34.34s %4d ", node->cn_name,
- node->cn_nodeid);
+ printf(fmt_buf, node->cn_name);
+ printf("%4d ", node->cn_nodeid);

if (node->cn_member & FLAG_UP)
printf("Online");
@@ -640,15 +738,21 @@
if (node->cn_member & FLAG_NOCFG)
printf(", Estranged");

- if (node->cn_member & FLAG_RGMGR)
- printf(", rgmanager");
+ if (node->cn_member & FLAG_RGMGR) {
+ if (rgmanager_master_present) {
+ if (node->cn_member & FLAG_RGMAST)
+ printf(", RG-Master");
+ else
+ printf(", RG-Worker");
+ } else {
+ printf(", rgmanager");
+ }
+ }

if (node->cn_member & FLAG_QDISK)
printf(", Quorum Disk");

printf("
");
-
-
}

@@ -656,12 +760,14 @@
xml_member_state(cman_node_t *node)
{
printf(" <node name="%s" state="%d" local="%d" "
- "estranged="%d" rgmanager="%d" qdisk="%d" nodeid="0x%08x"/>
",
+ "estranged="%d" rgmanager="%d" rgmanager_master="%d" "
+ "qdisk="%d" nodeid="0x%08x"/>
",
node->cn_name,
!!(node->cn_member & FLAG_UP),
!!(node->cn_member & FLAG_LOCAL),
!!(node->cn_member & FLAG_NOCFG),
!!(node->cn_member & FLAG_RGMGR),
+ !!(node->cn_member & FLAG_RGMAST),
!!(node->cn_member & FLAG_QDISK),
(uint32_t)((node->cn_nodeid )&0xffffffff));
}
@@ -670,6 +776,7 @@
int
txt_member_states(cluster_member_list_t *membership, char *name)
{
+ char buf[80];
int x, ret = 0;

if (!membership) {
@@ -677,13 +784,17 @@
return -1;
}

- printf(" %-34.34s %-4.4s %s
", "Member Name", "ID", "Status");
- printf(" %-34.34s %-4.4s %s
", "------ ----", "----", "------");
+ build_member_format(buf, sizeof(buf), dimx);
+
+ printf(buf, "Member Name");
+ printf("%-4.4s %s
", "ID", "Status");
+ printf(buf, "------ ----");
+ printf("%-4.4s %s
", "----", "------");

for (x = 0; x < membership->cml_count; x++) {
if (name && strcmp(membership->cml_members[x].cn_name, name))
continue;
- txt_member_state(&membership->cml_members[x]);
+ txt_member_state(&membership->cml_members[x], buf);
ret = !(membership->cml_members[x].cn_member & FLAG_UP);
}

@@ -717,13 +828,15 @@

int
-txt_cluster_status(int qs, cluster_member_list_t *membership,
+txt_cluster_status(cman_cluster_t *ci,
+ int qs, cluster_member_list_t *membership,
rg_state_list_t *rgs, char *name, char *svcname,
int flags)
{
int ret;

if (!svcname && !name) {
+ txt_cluster_info(ci);
txt_quorum_state(qs);
if (!membership) {
/* XXX Check for rgmanager?! */
@@ -738,12 +851,14 @@
return ret;
if (!name || (name && svcname))
ret = txt_rg_states(rgs, membership, svcname, flags);
+
return ret;
}

int
-xml_cluster_status(int qs, cluster_member_list_t *membership,
+xml_cluster_status(cman_cluster_t *ci, int qs,
+ cluster_member_list_t *membership,
rg_state_list_t *rgs, char *name, char *svcname,
int flags)
{
@@ -768,6 +883,8 @@
}

if (!svcname && !name)
+ xml_cluster_info(ci);
+ if (!svcname && !name)
xml_quorum_state(qs);
if (!svcname || (name && svcname))
ret1 = xml_member_states(membership, name);
@@ -862,7 +979,9 @@
int local_node_id;
int fast = 0;
int runtype = 0;
+ time_t now;
cman_handle_t ch = NULL;
+ cman_cluster_t ci;

int refresh_sec = 0, errors = 0;
int opt, xml = 0, flags = 0;
@@ -900,8 +1019,9 @@
case 's':
rg_name = optarg;
if (!strchr(rg_name,':')) {
+ memset(real_rg_name, 0, sizeof(real_rg_name));
snprintf(real_rg_name,
- sizeof(real_rg_name),
+ sizeof(real_rg_name)-1,
"service:%s", rg_name);
rg_name = real_rg_name;
}
@@ -941,7 +1061,7 @@
/* Connect & grab all our info */
ch = cman_init(NULL);
if (!ch) {
- printf("CMAN is not running.
");
+ perror("Could not connect to CMAN");
return 1;
}

@@ -952,8 +1072,12 @@
ret = !(cman_is_quorate(ch));
goto cleanup;
case VERSION_ONLY:
+#ifdef RELEASE_VERSION
printf("%s version %s
", basename(argv[0]),
RELEASE_VERSION);
+#else
+ printf("%s version DEVEL
", basename(argv[0]));
+#endif
if (!ch)
break;
goto cleanup;
@@ -974,6 +1098,16 @@
signal(SIGINT, term_handler);
signal(SIGTERM, term_handler);

+ if (isatty(STDOUT_FILENO)) {
+ stdout_is_tty = 1;
+ setupterm((char *) 0, STDOUT_FILENO, (int *) 0);
+ dimx = tigetnum("cols");
+ dimy = tigetnum("lines");
+ }
+
+ memset(&ci, 0, sizeof(ci));
+ cman_get_cluster(ch, &ci);
+
while (1) {
qs = cman_is_quorate(ch);
membership = build_member_list(ch, &local_node_id);
@@ -985,16 +1119,16 @@
}

if (refresh_sec) {
- setupterm((char *) 0, STDOUT_FILENO, (int *) 0);
tputs(clear_screen, lines > 0 ? lines : 1, putchar);
+ now = time(NULL);
}

if (xml)
- ret = xml_cluster_status(qs, membership, rgs,
+ ret = xml_cluster_status(&ci, qs, membership, rgs,
member_name, rg_name,
flags);
else
- ret = txt_cluster_status(qs, membership, rgs,
+ ret = txt_cluster_status(&ci, qs, membership, rgs,
member_name, rg_name,
flags);

@@ -1011,5 +1145,6 @@

cleanup:
cman_finish(ch);
+
return ret;
}
--- cluster/rgmanager/src/utils/clusvcadm.c 2007/08/22 08:58:47 1.22
+++ cluster/rgmanager/src/utils/clusvcadm.c 2007/11/30 20:06:55 1.23
@@ -311,7 +311,11 @@
node_specified = 1;
break;
case 'v':
- printf("%s
",RELEASE_VERSION);
+#ifdef RELEASE_VERSION
+ printf("%s
", RELEASE_VERSION);
+#else
+ printf("DEVEL
");
+#endif
return 0;
case 'Z':
actionstr = "freezing";