cluster-commits February 2010

cluster-commits@lists.fedorahosted.org

9 participants
62 discussions

cluster: STABLE3 - fence agents: fix several agents build
by Fabio M. Di Nitto 08 Feb '10

08 Feb '10

Gitweb: http://git.fedorahosted.org/git/cluster.git?p=cluster.git;a=commitdiff;h=b5… Commit: b5a948214b9573970122c8c4fc07acfb56f049fc Parent: 818d24bc8cea95cceb9310d7798b75cb8cb34cd6 Author: Fabio M. Di Nitto <fdinitto(a)redhat.com> AuthorDate: Mon Feb 8 13:54:09 2010 +0100 Committer: Fabio M. Di Nitto <fdinitto(a)redhat.com> CommitterDate: Mon Feb 8 13:54:09 2010 +0100 fence agents: fix several agents build Signed-off-by: Fabio M. Di Nitto <fdinitto(a)redhat.com> --- scripts/fenceparse | 4 +++- 1 files changed, 3 insertions(+), 1 deletions(-) diff --git a/scripts/fenceparse b/scripts/fenceparse index c34cb82..49457a6 100644 --- a/scripts/fenceparse +++ b/scripts/fenceparse @@ -10,7 +10,9 @@ infile="$5" definedata="$(cat $definefile | grep "^\#define $define" | sed -e 's/.*'$define' //')" -realinfile="$(ls $srcdir/$infile.*)" +realinfile="$(ls $srcdir/$infile.*{py,pl,sh} 2>/dev/null || true)" + +[ -z "$realinfile" ] && exit 1 interpreter="$(cat $realinfile | head -n 1 | awk -F "/" '{print $NF}')" interpreter="$(echo $interpreter)"

1 0

gfs2-utils: master - mount.gfs2: Better error reporting when mounting a gfs fs without enough journals
by Abhijith Das 05 Feb '10

05 Feb '10

Gitweb: http://git.fedorahosted.org/git/gfs2-utils.git?p=gfs2-utils.git;a=commitdif… Commit: 71f46b8688811540f0bfd53b63b55915a5b34ebc Parent: 06b5fb874d5bd782cf64d6f451098ea684ca6e67 Author: Abhijith Das <adas(a)redhat.com> AuthorDate: Fri Feb 5 17:05:53 2010 -0600 Committer: Abhijith Das <adas(a)redhat.com> CommitterDate: Fri Feb 5 17:05:53 2010 -0600 mount.gfs2: Better error reporting when mounting a gfs fs without enough journals use EUSERS to signal to mount helper about insufficient journals so it doesn't display a cryptic message rhbz#537201 --- gfs2/mount/mount.gfs2.c | 2 ++ 1 files changed, 2 insertions(+), 0 deletions(-) diff --git a/gfs2/mount/mount.gfs2.c b/gfs2/mount/mount.gfs2.c index c24d056..610a7f8 100644 --- a/gfs2/mount/mount.gfs2.c +++ b/gfs2/mount/mount.gfs2.c @@ -212,6 +212,8 @@ int main(int argc, char **argv) if (errno == EBUSY) die("%s already mounted or %s busy\n", mo.dev, mo.dir); + else if (errno == EUSERS) + die("Too many nodes mounting filesystem, no free journals\n"); die("error mounting %s on %s: %s\n", mo.dev, mo.dir, strerror(errno)); }

1 0

cluster: STABLE3 - Merge branch 'STABLE3' of ssh://git.fedoraproject.org/git/cluster into mySTABLE3
by Abhijith Das 05 Feb '10

05 Feb '10

Gitweb: http://git.fedorahosted.org/git/cluster.git?p=cluster.git;a=commitdiff;h=81… Commit: 818d24bc8cea95cceb9310d7798b75cb8cb34cd6 Parent: 6c62ef54d3d90d424c055cb950e0e3aef27a6601 6d2ce78aa8124a1586ed950bb7627f89f0f9d2b8 Author: Abhijith Das <adas(a)redhat.com> AuthorDate: Fri Feb 5 16:56:56 2010 -0600 Committer: Abhijith Das <adas(a)redhat.com> CommitterDate: Fri Feb 5 16:56:56 2010 -0600 Merge branch 'STABLE3' of ssh://git.fedoraproject.org/git/cluster into mySTABLE3 gfs2/init.d/gfs2.in | 155 +++++++++++++++++++++++---------------------------- 1 files changed, 69 insertions(+), 86 deletions(-)

1 0

cluster: STABLE3 - mount.gfs2: Better error reporting when mounting a gfs fs without enough journals
by Abhijith Das 05 Feb '10

05 Feb '10

Gitweb: http://git.fedorahosted.org/git/cluster.git?p=cluster.git;a=commitdiff;h=6c… Commit: 6c62ef54d3d90d424c055cb950e0e3aef27a6601 Parent: b55c086de3d0eba532f8ced97da92d1117718b22 Author: Abhijith Das <adas(a)redhat.com> AuthorDate: Fri Feb 5 16:53:37 2010 -0600 Committer: Abhijith Das <adas(a)redhat.com> CommitterDate: Fri Feb 5 16:53:37 2010 -0600 mount.gfs2: Better error reporting when mounting a gfs fs without enough journals use EUSERS to signal to mount helper about insufficient journals so it doesn't display a cryptic message rhbz#537201 --- gfs-kernel/src/gfs/ops_fstype.c | 2 +- gfs2/mount/mount.gfs2.c | 2 ++ 2 files changed, 3 insertions(+), 1 deletions(-) diff --git a/gfs-kernel/src/gfs/ops_fstype.c b/gfs-kernel/src/gfs/ops_fstype.c index 2e74d04..8bc76bf 100644 --- a/gfs-kernel/src/gfs/ops_fstype.c +++ b/gfs-kernel/src/gfs/ops_fstype.c @@ -437,7 +437,7 @@ static int init_journal(struct gfs_sbd *sdp, int undo) else { /* Discover this node's journal number (lock module tells us which one to use), and lock it */ - error = -EINVAL; + error = -EUSERS; if (sdp->sd_lockstruct.ls_jid >= sdp->sd_journals) { printk("GFS: fsid=%s: can't mount journal #%u\n", sdp->sd_fsname, sdp->sd_lockstruct.ls_jid); diff --git a/gfs2/mount/mount.gfs2.c b/gfs2/mount/mount.gfs2.c index 2442f59..38ec9a1 100644 --- a/gfs2/mount/mount.gfs2.c +++ b/gfs2/mount/mount.gfs2.c @@ -210,6 +210,8 @@ int main(int argc, char **argv) if (errno == EBUSY) die("%s already mounted or %s busy\n", mo.dev, mo.dir); + else if (errno == EUSERS) + die("Too many nodes mounting filesystem, no free journals\n"); die("error mounting %s on %s: %s\n", mo.dev, mo.dir, strerror(errno)); }

1 0

cluster: RHEL48 - fence_ilo: Add power-wait option to solve ilo2 v1.70 problem
by Marek Grác 05 Feb '10

05 Feb '10

Gitweb: http://git.fedorahosted.org/git/cluster.git?p=cluster.git;a=commitdiff;h=35… Commit: 35f94eace1e0b3b1c4b9da50d26ba8a1fdbf0716 Parent: 4f83d6506976ea57f97eb776aa22e46af15edd9d Author: Marek 'marx' Grac <mgrac(a)redhat.com> AuthorDate: Wed Jan 20 18:56:11 2010 +0100 Committer: Marek 'marx' Grac <mgrac(a)redhat.com> CommitterDate: Fri Feb 5 18:53:15 2010 +0100 fence_ilo: Add power-wait option to solve ilo2 v1.70 problem Default behaviour was not changed, it is possible to use -G <seconds>. ilo2 v1.70+ problem can be solved with '-G 5' rhbz#545682 --- fence/agents/ilo/fence_ilo.py | 2 +- fence/agents/lib/fencing.py.py | 6 ++++++ 2 files changed, 7 insertions(+), 1 deletions(-) diff --git a/fence/agents/ilo/fence_ilo.py b/fence/agents/ilo/fence_ilo.py index 6513e2a..8a34fb0 100755 --- a/fence/agents/ilo/fence_ilo.py +++ b/fence/agents/ilo/fence_ilo.py @@ -57,7 +57,7 @@ def set_power_status(conn, options): def main(): device_opt = [ "help", "version", "agent", "quiet", "verbose", "debug", "action", "ipaddr", "login", "passwd", "passwd_script", - "ssl", "ribcl" ] + "ssl", "ribcl", "power_wait" ] options = check_input(device_opt, process_input(device_opt)) diff --git a/fence/agents/lib/fencing.py.py b/fence/agents/lib/fencing.py.py index 461edb2..6e0e99c 100644 --- a/fence/agents/lib/fencing.py.py +++ b/fence/agents/lib/fencing.py.py @@ -137,6 +137,10 @@ all_opt = { "getopt" : "s:", "help" : "-s <id> Name of the managed system", "order" : 1 }, + "power_wait" : { + "getopt" : "G:", + "help" : "-G <seconds> Wait X seconds after issuing ON/OFF", + "order" : 1 }, "test" : { "getopt" : "T", "help" : "", @@ -360,6 +364,8 @@ def fence_action(tn, options, set_power_fn, get_power_fn): set_power_fn(tn, options) if wait_power_status(tn, options, get_power_fn) == 0: fail(EC_WAITING_OFF) + if options.has_key("-G"): + time.sleep(int(options["-G"])) options["-o"] = "on" set_power_fn(tn, options) if wait_power_status(tn, options, get_power_fn) == 0:

1 0

gfs2-utils: master - gfs2: man page updates
by Steven Whitehouse 04 Feb '10

04 Feb '10

Gitweb: http://git.fedorahosted.org/git/gfs2-utils.git?p=gfs2-utils.git;a=commitdif… Commit: 06b5fb874d5bd782cf64d6f451098ea684ca6e67 Parent: f581605e0a8bf8eeacbedaddfa134924b221cced Author: Steven Whitehouse <swhiteho(a)redhat.com> AuthorDate: Thu Feb 4 14:34:57 2010 +0000 Committer: Steven Whitehouse <swhiteho(a)redhat.com> CommitterDate: Thu Feb 4 14:34:57 2010 +0000 gfs2: man page updates This brings the man pages uptodate. The main change is to document the current set of mount options and to remove those which are obsolete. Signed-off-by: Steven Whitehouse <swhiteho(a)redhat.com> --- gfs2/man/gfs2_tool.8 | 12 ++++- gfs2/man/mount.gfs2.8 | 108 ++++++++++++++++++++++++++++++++++--------------- 2 files changed, 84 insertions(+), 36 deletions(-) diff --git a/gfs2/man/gfs2_tool.8 b/gfs2/man/gfs2_tool.8 index 1533de7..dccd631 100644 --- a/gfs2/man/gfs2_tool.8 +++ b/gfs2/man/gfs2_tool.8 @@ -16,7 +16,10 @@ as gfs_tool used to. .SH COMMANDS .TP \fBclearflag\fP \fIFlag\fR \fIFile1\fR \fIFile2\fR \fI...\fR -Clear an attribute flag on a file. See \fBsetflag\fP for available flags. +Clear an attribute flag on a file. This is now obsolete and kept +only for backward compatibility, chattr is the preferred way to +clear attribute flags. See \fBsetflag\fP for available flags. This +option will probably be removed at a future date. .TP \fBfreeze\fP \fIMountPoint\fR Freeze (quiesce) a GFS2 cluster. @@ -35,7 +38,8 @@ Print out information about the journals in a mounted filesystem. \fBlockdump\fP \fIMountPoint\fR Print out information about the locks this machine holds for a given filesystem. This information is also available via the debugfs -glock dump file. +glock dump file, and accessing that file is the preferred method +of obtaining a dump of the glock state. .\".TP .\"\fBrindex\fP \fIMountPoint\fR .\"Print out the resource group index of a mounted filesystem. @@ -73,7 +77,7 @@ Print out the superblock. Set an attribute flag on a file. The currently supported flags are jdata, immutable, appendonly, noatime, and sync. The chattr command is the preferred way to set attributes on -files. +files. This option will probably be removed at a future date. The \fIjdata\fR flag causes all the data written to a file to be journaled. If the \fIjdata\fR flag is set for a directory, @@ -108,4 +112,6 @@ Print out the version of GFS2 that this program goes with. .TP \fBwithdraw\fP \fIMountPoint\fR Cause GFS2 to abnormally shutdown a given filesystem on this node. +This feature is only useful for testing and should not be used +during normal filesystem operation. diff --git a/gfs2/man/mount.gfs2.8 b/gfs2/man/mount.gfs2.8 index fcf99c4..29d14c8 100644 --- a/gfs2/man/mount.gfs2.8 +++ b/gfs2/man/mount.gfs2.8 @@ -4,18 +4,17 @@ mount.gfs2 - GFS2 mount options .SH SYNOPSIS -.B mount -[\fIStandardMountOptions\fR] \fB-t\fP gfs2 \fIDEVICE\fR \fIMOUNTPOINT\fR \fB-o\fP [GFS2Option1,GFS2Option2,GFS2OptionX...] +.BI "mount \-a [\-fnrsvw] \-t " gfs2 " [\-O " options "] +.br +.BI "mount [\-fnrsvw] \-t " gfs2 " [\-o " options " ] " "device dir" .SH DESCRIPTION -GFS2 may be used as a local (single computer) filesystem, but its real purpose -is in clusters, where multiple computers (nodes) share a common storage device. - -Above is the format typically used to mount a GFS2 filesystem, using the -\fBmount\fP(8) command. The \fIdevice\fR may be any block device on which you +For details on the common mount options, please see the +\fBmount\fP(8) command man page. +The \fIdevice\fR may be any block device on which you have created a GFS2 filesystem. Examples include a single disk partition (e.g. /dev/sdb3), a loopback device, a device exported -from another node (e.g. an iSCSI device or a \fBgnbd\fP(8) device), or a +from another node (e.g. an iSCSI device), or a logical volume (typically comprised of a number of individual disks). \fIdevice\fR does not necessarily need to match the device name as seen on @@ -31,8 +30,6 @@ different computers. Each computer must see an entire filesystem. You may, however, mount several GFS2 filesystems if you want to distribute your data storage in a controllable way. -\fImountpoint\fR is the same as \fIdir\fR in the \fBmount\fP(8) man page. - This man page describes GFS2-specific options that can be passed to the GFS2 file system at mount time, using the \fB-o\fP flag. There are many other \fB-o\fP options handled by the generic mount command \fBmount\fP(8). @@ -41,9 +38,9 @@ interpreted by the mount command nor by the kernel's Virtual File System. GFS2 and non-GFS2 options may be intermingled after the \fB-o\fP, separated by commas (but no spaces). -The options debug, commit, discard, acl, quota, suiddir, and data can be +The options commit, discard, errors, quota_quantum, statfs_quantum, statfs_percent, barrier, acl, quota, suiddir, and data can be changed after mount using the "mount -o remount,option /mountpoint" command. -The options debug, quota, discard, acl, and suiddir support the "no" +The options quota, discard, barrier, acl, and suiddir support the "no" prefix. For example, "noacl" turns off what "acl" turns on. If you have trouble mounting GFS2, check the syslog (e.g. /var/log/messages) @@ -51,15 +48,13 @@ for specific error messages. .SH OPTIONS .TP -\fBlockproto=\fP\fILockModuleName\fR +\fBlockproto=\fP\fILockProtoName\fR This specifies which inter-node lock protocol is used by the GFS2 filesystem for this mount, overriding the default lock protocol name stored in the filesystem's on-disk superblock. -The \fILockModuleName\fR must be an exact match of the protocol name presented -by the lock module when it registers with the lock harness. Traditionally, -this matches the .o filename of the lock module, e.g. \fIlock_dlm\fR, -or \fIlock_nolock\fR. +The \fILockProtoName\fR must be one of the supported locking protocols, +currently these are \fIlock_nolock\fR and \fIlock_dlm\fR. The default lock protocol name is written to disk initially when creating the filesystem with \fBmkfs.gfs2\fP(8), -p option. It can be changed on-disk by @@ -67,7 +62,9 @@ using the \fBgfs2_tool\fP(8) utility's \fBsb proto\fP command. The \fBlockproto\fP mount option should be used only under special circumstances in which you want to temporarily use a different lock protocol -without changing the on-disk default. +without changing the on-disk default. Using the incorrect lock protocol +on a cluster filesystem mounted from more than one node will almost +certainly result in filesystem corruption. .TP \fBlocktable=\fP\fILockTableName\fR This specifies the identity of the cluster and of the filesystem for this @@ -78,8 +75,8 @@ the inter-node locking system, enabling the mounting of multiple GFS2 filesystems. The format of \fILockTableName\fR is lock-module-specific. For -lock_dlm, the format is \fIclustername:fsname\fR. For -lock_nolock, the field is ignored. +\fIlock_dlm\fR, the format is \fIclustername:fsname\fR. For +\fIlock_nolock\fR, the field is ignored. The default cluster/filesystem name is written to disk initially when creating the filesystem with \fBmkfs.gfs2\fP(8), -t option. It can be changed on-disk @@ -108,10 +105,13 @@ is achieved by letting VFS handle the whole job. This is turned on automatically by the lock_nolock module, but can be overridden by using the \fBignore_local_fs\fP option. .TP -\fBdebug\fP -Causes GFS2 to oops when encountering an error that would cause the -mount to withdraw or print an assertion warning. This option should -probably not be used in a production system. +\fBerrors=\fP\fI[panic|withdraw]\fR +Setting errors=panic causes GFS2 to oops when encountering an error that +would otherwise cause the +mount to withdraw or print an assertion warning. The default setting +is errors=withdraw. This option should not be used in a production system. +It replaces the earlier \fBdebug\fP option on kernel versions 2.6.31 and +above. .TP \fBignore_local_fs\fP By default, using the nolock lock module automatically turns on the @@ -133,7 +133,8 @@ Enables POSIX Access Control List \fBacl\fP(5) support within GFS2. .TP \fBspectator\fP Mount this filesystem using a special form of read-only mount. The mount -does not use one of the filesystem's journals. +does not use one of the filesystem's journals. The node is unable to +recover journals for other nodes. .TP \fBsuiddir\fP Sets owner of any newly created file or directory to be that of parent @@ -155,13 +156,21 @@ been freed. These can be used by suitable hardware to implement thin-provisioning and similar schemes. This feature is supported in kernel version 2.6.30 and above. .TP +\fBbarrier\fP +This option, which defaults to on, causes GFS2 to send I/O barriers +when flushing the journal. The option is automatically turned off +if the underlying device does not support I/O barriers. We highly +recommend the use of I/O barriers with GFS2 at all times unless +the block device is designed so that it cannot lose its write cache +content (e.g. its on a UPS, or it doesn't have a write cache) +.TP \fBcommit=\fP\fIsecs\fR This is similar to the ext3 \fBcommit=\fP option in that it sets the maximum number of seconds between journal commits if there is dirty data in the journal. The default is 60 seconds. This option is only provided in kernel versions 2.6.31 and above. .TP -\fBdata=\fP\fI[ordered/writeback]\fR +\fBdata=\fP\fI[ordered|writeback]\fR When data=ordered is set, the user data modified by a transaction is flushed to the disk before the transaction is committed to disk. This should prevent the user from seeing uninitialized blocks in a file @@ -169,14 +178,47 @@ after a crash. Data=writeback mode writes the user data to the disk at any time after it's dirtied. This doesn't provide the same consistency guarantee as ordered mode, but it should be slightly faster for some workloads. The default is ordered mode. - -.SH LINKS -.TP 30 -http://sources.redhat.com/cluster --- home site of GFS2 .TP -http://www.suse.de/~agruen/acl/linux-acls/ --- good writeup on ACL support in Linux +\fBmeta\fP +This option results in selecting the meta filesystem root rather than +the normal filesystem root. This option is normally only used by +the GFS2 utility functions. Altering any file on the GFS2 meta filesystem +may render the filesystem unusable, so only experts in the GFS2 +on-disk layout should use this option. +.TP +\fBquota_quantum=\fP\fIsecs\fR +This sets the number of seconds for which a change in the quota +information may sit on one node before being written to the quota +file. This is the preferred way to set this parameter. The value +is an integer number of seconds greater than zero. The default is +60 seconds. Shorter settings result in faster updates of the lazy +quota information and less likelihood of someone exceeding their +quota. Longer settings make filesystem operations involving quotas +faster and more efficient. +.TP +\fBstatfs_quantum=\fP\fIsecs\fR +Setting statfs_quantum to 0 is the preferred way to set the slow version +of statfs. The default value is 30 secs which sets the maximum time +period before statfs changes will be syned to the master statfs file. +This can be adjusted to allow for faster, less accurate statfs values +or slower more accurate values. When set to 0, statfs will always +report the true values. +.TP +\fBstatfs_percent=\fP\fIvalue\fR +This setting provides a bound on the maximum percentage change in +the statfs information on a local basis before it is synced back +to the master statfs file, even if the time period has not +expired. If the setting of statfs_quantum is 0, then this setting +is ignored. + +.SH BUGS + +GFS2 doesn't support \fBerrors=\fP\fIremount-ro\fR or \fBdata=\fP\fIjournal\fR. +It is not possible to switch support for user and group quotas on and +off independently of each other. Some of the error messages are rather +cryptic, if you encounter one of these messages check firstly that gfs_controld +is running and secondly that you have enough journals on the filesystem +for the number of nodes in use. .SH SEE ALSO

1 0

gfs2-utils: master - gfs2: make init script LSB compliant
by Fabio M. Di Nitto 03 Feb '10

03 Feb '10

Gitweb: http://git.fedorahosted.org/git/gfs2-utils.git?p=gfs2-utils.git;a=commitdif… Commit: f581605e0a8bf8eeacbedaddfa134924b221cced Parent: e9fa3a0ae363f8bca7a89aaebf1f46177670321e Author: Fabio M. Di Nitto <fdinitto(a)redhat.com> AuthorDate: Wed Feb 3 14:41:55 2010 +0100 Committer: Fabio M. Di Nitto <fdinitto(a)redhat.com> CommitterDate: Wed Feb 3 14:48:06 2010 +0100 gfs2: make init script LSB compliant clean up white spaces remove lots of unnecessary and/or obsolete bits rhbz#536962 rhbz#553383 Signed-off-by: Fabio M. Di Nitto <fdinitto(a)redhat.com> --- gfs2/init.d/gfs2.in | 155 +++++++++++++++++++++++---------------------------- 1 files changed, 69 insertions(+), 86 deletions(-) diff --git a/gfs2/init.d/gfs2.in b/gfs2/init.d/gfs2.in index 241e3f9..e957a36 100644 --- a/gfs2/init.d/gfs2.in +++ b/gfs2/init.d/gfs2.in @@ -18,14 +18,28 @@ # set secure PATH PATH="/bin:/usr/bin:/sbin:/usr/sbin:@SBINDIR@" +### generic wrapper functions + success() { - echo -ne "[ OK ]\r" + echo -ne "[ OK ]\r" } failure() { - echo -ne "[FAILED]\r" + echo -ne "[FAILED]\r" +} + +ok() { + success + echo +} + +nok() { + echo -e "$errmsg" + failure + echo + exit 1 } # rpm based distros @@ -43,106 +57,75 @@ if [ -d /etc/default ]; then [ -z "$LOCK_FILE" ] && LOCK_FILE="/var/lock/gfs2" fi +# proc is required for both status and stop. +# start could live without, but better be consistent with the behavior +if [ ! -f /proc/mounts ]; then + echo "GFS2: /proc is not available, unable to proceed" + exit 1 +fi + # # This script's behavior is modeled closely after the netfs script. # GFS2FSTAB=$(LC_ALL=C awk '!/^#/ && $3 == "gfs2" && $4 !~ /noauto/ { print $2 }' /etc/fstab) GFS2MTAB=$(LC_ALL=C awk '!/^#/ && $3 == "gfs2" && $2 != "/" { print $2 }' /proc/mounts) +if [ -z "$GFS2FSTAB" ]; then + echo "GFS2: no entries found in /etc/fstab" + exit 6 +fi + # See how we were called. case "$1" in - start) - if [ -n "$GFS2FSTAB" ] - then - echo -n "Mounting GFS2 filesystems: " - mount -a -t gfs2 - rtrn=$? - if [ $rtrn = 0 ]; then - touch $LOCK_FILE - success - echo - else - failure - echo - fi - fi +start) + [ -z "$GFS2FSTAB" ] && exit 0 + echo -n "Mounting GFS2 filesystems: " + errmsg="$(mount -a -t gfs2 2>&1)" || nok + touch $LOCK_FILE + ok +;; +stop) + [ -z "$GFS2MTAB" ] && exit 0 + echo -n "Unmounting GFS2 filesystems: " + errmsg="$(umount -a -t gfs2 2>&1)" || nok + modprobe -r gfs2 > /dev/null 2>&1 || true + rm -f $LOCK_FILE + ok ;; - stop) - if [ -n "$GFS2MTAB" ] - then - sig= - retry=6 - remaining=`LC_ALL=C awk '!/^#/ && $3 == "gfs2" && $2 != "/" {print $2}' /proc/mounts` - while [ -n "$remaining" -a "$retry" -gt 0 ] - do - echo -n "Unmounting GFS2 filesystems: " - umount -a -t gfs2 - rtrn=$? - if [ $rtrn = 0 ]; then - success - echo - else - failure - echo - fi - - if [ $retry -eq 0 ] - then - echo -n "Unmounting GFS2 filesystems (lazy): " - umount -l -a -t gfs2 - rtrn=$? - if [ $rtrn = 0 ]; then - success - echo - else - failure - echo - fi - break - fi - - sleep 2 - remaining=`LC_ALL=C awk '!/^#/ && $3 == "gfs2" && $2 != "/" {print $2}' /proc/mounts` - [ -z "$remaining" ] && break - fuser -k -m $sig $remaining > /dev/null 2>&1 - sleep 10 - retry=$(($retry - 1)) - sig=-9 - done +status) + if [ -z "$GFS2MTAB" ] && [ -f $LOCK_FILE ]; then + echo "GFS2: Found stale lock file $LOCK_FILE" + exit 2 fi - modprobe -r gfs2 - rm -f $LOCK_FILE - ;; - - status) - if [ -f /proc/mounts ] - then - [ -n "$GFS2FSTAB" ] && { - echo "Configured GFS2 mountpoints: " - for fs in $GFS2FSTAB; do echo $fs ; done - } - [ -n "$GFS2MTAB" ] && { - echo "Active GFS2 mountpoints: " - for fs in $GFS2MTAB; do echo $fs ; done - } - else - echo "/proc filesystem unavailable" + if [ -n "$GFS2FSTAB" ] && [ -z "$GFS2MTAB" ]; then + echo "GFS2: service is not running" + exit 3 fi - ;; - restart) + echo "Configured GFS2 mountpoints: " + for fs in $GFS2FSTAB; do + echo $fs; + done + + echo "Active GFS2 mountpoints: " + for fs in $GFS2MTAB; do + echo $fs; + done +;; +condrestart|try-restart) + $0 status >/dev/null 2>&1 || exit 0 + $0 restart +;; +restart|reload|force-reload) $0 stop $0 start - ;; - - reload) - $0 start - ;; - *) - echo $"Usage: $0 {start|stop|restart|reload|status}" - exit 1 +;; +*) + echo "Usage: $0 {start|stop|restart|reload|force-reload|condrestart|try-restart|status}" + exit 2 +;; esac exit 0

1 0

cluster: STABLE3 - gfs2: make init script LSB compliant
by Fabio M. Di Nitto 03 Feb '10

03 Feb '10

Gitweb: http://git.fedorahosted.org/git/cluster.git?p=cluster.git;a=commitdiff;h=6d… Commit: 6d2ce78aa8124a1586ed950bb7627f89f0f9d2b8 Parent: b55c086de3d0eba532f8ced97da92d1117718b22 Author: Fabio M. Di Nitto <fdinitto(a)redhat.com> AuthorDate: Wed Feb 3 14:41:55 2010 +0100 Committer: Fabio M. Di Nitto <fdinitto(a)redhat.com> CommitterDate: Wed Feb 3 14:41:55 2010 +0100 gfs2: make init script LSB compliant clean up white spaces remove lots of unnecessary and/or obsolete bits rhbz#536962 rhbz#553383 Signed-off-by: Fabio M. Di Nitto <fdinitto(a)redhat.com> --- gfs2/init.d/gfs2.in | 155 +++++++++++++++++++++++---------------------------- 1 files changed, 69 insertions(+), 86 deletions(-) diff --git a/gfs2/init.d/gfs2.in b/gfs2/init.d/gfs2.in index b28a99a..e957a36 100644 --- a/gfs2/init.d/gfs2.in +++ b/gfs2/init.d/gfs2.in @@ -18,14 +18,28 @@ # set secure PATH PATH="/bin:/usr/bin:/sbin:/usr/sbin:@SBINDIR@" +### generic wrapper functions + success() { - echo -ne "[ OK ]\r" + echo -ne "[ OK ]\r" } failure() { - echo -ne "[FAILED]\r" + echo -ne "[FAILED]\r" +} + +ok() { + success + echo +} + +nok() { + echo -e "$errmsg" + failure + echo + exit 1 } # rpm based distros @@ -43,106 +57,75 @@ if [ -d /etc/default ]; then [ -z "$LOCK_FILE" ] && LOCK_FILE="/var/lock/gfs2" fi +# proc is required for both status and stop. +# start could live without, but better be consistent with the behavior +if [ ! -f /proc/mounts ]; then + echo "GFS2: /proc is not available, unable to proceed" + exit 1 +fi + # # This script's behavior is modeled closely after the netfs script. # GFS2FSTAB=$(LC_ALL=C awk '!/^#/ && $3 == "gfs2" && $4 !~ /noauto/ { print $2 }' /etc/fstab) GFS2MTAB=$(LC_ALL=C awk '!/^#/ && $3 == "gfs2" && $2 != "/" { print $2 }' /proc/mounts) +if [ -z "$GFS2FSTAB" ]; then + echo "GFS2: no entries found in /etc/fstab" + exit 6 +fi + # See how we were called. case "$1" in - start) - if [ -n "$GFS2FSTAB" ] - then - echo -n "Mounting GFS2 filesystems: " - mount -a -t gfs2 - rtrn=$? - if [ $rtrn = 0 ]; then - touch $LOCK_FILE - success - echo - else - failure - echo - fi - fi +start) + [ -z "$GFS2FSTAB" ] && exit 0 + echo -n "Mounting GFS2 filesystems: " + errmsg="$(mount -a -t gfs2 2>&1)" || nok + touch $LOCK_FILE + ok +;; +stop) + [ -z "$GFS2MTAB" ] && exit 0 + echo -n "Unmounting GFS2 filesystems: " + errmsg="$(umount -a -t gfs2 2>&1)" || nok + modprobe -r gfs2 > /dev/null 2>&1 || true + rm -f $LOCK_FILE + ok ;; - stop) - if [ -n "$GFS2MTAB" ] - then - sig= - retry=6 - remaining=`LC_ALL=C awk '!/^#/ && $3 == "gfs2" && $2 != "/" {print $2}' /proc/mounts` - while [ -n "$remaining" -a "$retry" -gt 0 ] - do - echo -n "Unmounting GFS2 filesystems: " - umount -a -t gfs2 - rtrn=$? - if [ $rtrn = 0 ]; then - success - echo - else - failure - echo - fi - - if [ $retry -eq 0 ] - then - echo -n "Unmounting GFS2 filesystems (lazy): " - umount -l -a -t gfs2 - rtrn=$? - if [ $rtrn = 0 ]; then - success - echo - else - failure - echo - fi - break - fi - - sleep 2 - remaining=`LC_ALL=C awk '!/^#/ && $3 == "gfs2" && $2 != "/" {print $2}' /proc/mounts` - [ -z "$remaining" ] && break - fuser -k -m $sig $remaining > /dev/null 2>&1 - sleep 10 - retry=$(($retry - 1)) - sig=-9 - done +status) + if [ -z "$GFS2MTAB" ] && [ -f $LOCK_FILE ]; then + echo "GFS2: Found stale lock file $LOCK_FILE" + exit 2 fi - modprobe -r gfs2 - rm -f $LOCK_FILE - ;; - - status) - if [ -f /proc/mounts ] - then - [ -n "$GFS2FSTAB" ] && { - echo "Configured GFS2 mountpoints: " - for fs in $GFS2FSTAB; do echo $fs ; done - } - [ -n "$GFS2MTAB" ] && { - echo "Active GFS2 mountpoints: " - for fs in $GFS2MTAB; do echo $fs ; done - } - else - echo "/proc filesystem unavailable" + if [ -n "$GFS2FSTAB" ] && [ -z "$GFS2MTAB" ]; then + echo "GFS2: service is not running" + exit 3 fi - ;; - restart) + echo "Configured GFS2 mountpoints: " + for fs in $GFS2FSTAB; do + echo $fs; + done + + echo "Active GFS2 mountpoints: " + for fs in $GFS2MTAB; do + echo $fs; + done +;; +condrestart|try-restart) + $0 status >/dev/null 2>&1 || exit 0 + $0 restart +;; +restart|reload|force-reload) $0 stop $0 start - ;; - - reload|force-reload) - $0 start - ;; - *) - echo $"Usage: $0 {start|stop|restart|reload|status}" - exit 1 +;; +*) + echo "Usage: $0 {start|stop|restart|reload|force-reload|condrestart|try-restart|status}" + exit 2 +;; esac exit 0

1 0

gfs2-utils: master - gfs2_convert: gfs2_convert should fix statfs file
by Abhijith Das 02 Feb '10

02 Feb '10

Gitweb: http://git.fedorahosted.org/git/gfs2-utils.git?p=gfs2-utils.git;a=commitdif… Commit: e9fa3a0ae363f8bca7a89aaebf1f46177670321e Parent: 23cdad0c9884594339cc2adb150356c48bd83ab0 Author: Abhijith Das <adas(a)redhat.com> AuthorDate: Tue Feb 2 14:29:15 2010 -0600 Committer: Abhijith Das <adas(a)redhat.com> CommitterDate: Tue Feb 2 14:38:35 2010 -0600 gfs2_convert: gfs2_convert should fix statfs file This patch modifies the freedi() code path in libgfs2 to update the blks_alloced and dinodes_alloced counts in the superblock struct. It also moves some code around in gfs2_convert so that all the block changes are correctly recorded into the statfs file. rhbz#556961 --- gfs2/convert/gfs2_convert.c | 9 +++++---- gfs2/libgfs2/fs_ops.c | 5 +++++ 2 files changed, 10 insertions(+), 4 deletions(-) diff --git a/gfs2/convert/gfs2_convert.c b/gfs2/convert/gfs2_convert.c index 046c9da..3bd42ad 100644 --- a/gfs2/convert/gfs2_convert.c +++ b/gfs2/convert/gfs2_convert.c @@ -1353,10 +1353,10 @@ static int journ_space_to_rg(struct gfs2_sbd *sdp) rgd->ri.ri_data0 = jndx->ji_addr + rgd->ri.ri_length; rgd->ri.ri_data = size - rgd->ri.ri_length; - sdp->blks_total += rgd->ri.ri_data; /* For statfs file update */ /* Round down to nearest multiple of GFS2_NBBY */ while (rgd->ri.ri_data & 0x03) rgd->ri.ri_data--; + sdp->blks_total += rgd->ri.ri_data; /* For statfs file update */ rgd->rg.rg_free = rgd->ri.ri_data; rgd->ri.ri_bitbytes = rgd->ri.ri_data / GFS2_NBBY; @@ -1417,7 +1417,7 @@ static void write_statfs_file(struct gfs2_sbd *sdp) struct gfs2_statfs_change sc; char buf[sizeof(struct gfs2_statfs_change)]; int count; - + sc.sc_total = sdp->blks_total; sc.sc_free = sdp->blks_total - sdp->blks_alloced; sc.sc_dinodes = sdp->dinodes_alloced; @@ -1588,6 +1588,9 @@ int main(int argc, char **argv) build_quota(&sb2); update_inode_file(&sb2); + /* Now delete the now-obsolete gfs1 files: */ + remove_obsolete_gfs1(&sb2); + write_statfs_file(&sb2); inode_put(&sb2.master_dir); @@ -1596,8 +1599,6 @@ int main(int argc, char **argv) fsync(sb2.device_fd); /* write the buffers to disk */ - /* Now delete the now-obsolete gfs1 files: */ - remove_obsolete_gfs1(&sb2); /* Now free all the in memory */ gfs2_rgrp_free(&sb2.rglist); log_notice("Committing changes to disk.\n"); diff --git a/gfs2/libgfs2/fs_ops.c b/gfs2/libgfs2/fs_ops.c index 56cd4c4..4a6d7e9 100644 --- a/gfs2/libgfs2/fs_ops.c +++ b/gfs2/libgfs2/fs_ops.c @@ -1675,6 +1675,7 @@ void gfs2_free_block(struct gfs2_sbd *sdp, uint64_t block) gfs2_set_bitmap(sdp, block, GFS2_BLKST_FREE); rgd->rg.rg_free++; /* adjust the free count */ gfs2_rgrp_out(&rgd->rg, rgd->bh[0]); /* back to the buffer */ + sdp->blks_alloced--; } } @@ -1730,10 +1731,14 @@ int gfs2_freedi(struct gfs2_sbd *sdp, uint64_t diblock) /* Set the bitmap type for inode to free space: */ gfs2_set_bitmap(sdp, ip->i_di.di_num.no_addr, GFS2_BLKST_FREE); inode_put(&ip); + /* inode_put deallocated the extra block used by the dist inode, */ + /* so adjust it in the superblock struct */ + sdp->blks_alloced--; /* Now we have to adjust the rg freespace count and inode count: */ rgd = gfs2_blk2rgrpd(sdp, diblock); rgd->rg.rg_free++; rgd->rg.rg_dinodes--; /* one less inode in use */ gfs2_rgrp_out(&rgd->rg, rgd->bh[0]); + sdp->dinodes_alloced--; return 0; }

1 0

cluster: RHEL55 - gfs2_convert: gfs2_convert should fix statfs file
by Abhijith Das 02 Feb '10

02 Feb '10

Gitweb: http://git.fedorahosted.org/git/cluster.git?p=cluster.git;a=commitdiff;h=af… Commit: af9b2042ec0f0ca80229051bc158f18c241347ed Parent: 112a4979a64dc0e8cc74e5e63ae55b5788cf334e Author: Abhijith Das <adas(a)redhat.com> AuthorDate: Tue Feb 2 14:34:23 2010 -0600 Committer: Abhijith Das <adas(a)redhat.com> CommitterDate: Tue Feb 2 14:34:23 2010 -0600 gfs2_convert: gfs2_convert should fix statfs file This patch modifies the freedi() code path in libgfs2 to update the blks_alloced and dinodes_alloced counts in the superblock struct. It also moves some code around in gfs2_convert so that all the block changes are correctly recorded into the statfs file. rhbz#556961 --- gfs2/convert/gfs2_convert.c | 9 +++++---- gfs2/libgfs2/fs_ops.c | 5 +++++ 2 files changed, 10 insertions(+), 4 deletions(-) diff --git a/gfs2/convert/gfs2_convert.c b/gfs2/convert/gfs2_convert.c index cea8496..aa039ef 100644 --- a/gfs2/convert/gfs2_convert.c +++ b/gfs2/convert/gfs2_convert.c @@ -1367,10 +1367,10 @@ int journ_space_to_rg(struct gfs2_sbd *sdp) rgd->ri.ri_data0 = jndx->ji_addr + rgd->ri.ri_length; rgd->ri.ri_data = size - rgd->ri.ri_length; - sdp->blks_total += rgd->ri.ri_data; /* For statfs file update */ /* Round down to nearest multiple of GFS2_NBBY */ while (rgd->ri.ri_data & 0x03) rgd->ri.ri_data--; + sdp->blks_total += rgd->ri.ri_data; /* For statfs file update */ rgd->rg.rg_free = rgd->ri.ri_data; rgd->ri.ri_bitbytes = rgd->ri.ri_data / GFS2_NBBY; convert_bitmaps(sdp, rgd, FALSE); /* allocates rgd->bh */ @@ -1414,7 +1414,7 @@ void write_statfs_file(struct gfs2_sbd *sdp) struct gfs2_statfs_change sc; char buf[sizeof(struct gfs2_statfs_change)]; int count; - + sc.sc_total = sdp->blks_total; sc.sc_free = sdp->blks_total - sdp->blks_alloced; sc.sc_dinodes = sdp->dinodes_alloced; @@ -1581,6 +1581,9 @@ int main(int argc, char **argv) build_quota(&sb2); update_inode_file(&sb2); + /* Now delete the now-obsolete gfs1 files: */ + remove_obsolete_gfs1(&sb2); + write_statfs_file(&sb2); inode_put(sb2.master_dir, updated); @@ -1590,8 +1593,6 @@ int main(int argc, char **argv) bcommit(&sb2.buf_list); /* write the buffers to disk */ bcommit(&sb2.nvbuf_list); /* write the buffers to disk */ - /* Now delete the now-obsolete gfs1 files: */ - remove_obsolete_gfs1(&sb2); /* Now free all the in memory */ gfs2_rgrp_free(&sb2.rglist, updated); log_notice("Committing changes to disk.\n"); diff --git a/gfs2/libgfs2/fs_ops.c b/gfs2/libgfs2/fs_ops.c index 9cc6dea..675dcd6 100644 --- a/gfs2/libgfs2/fs_ops.c +++ b/gfs2/libgfs2/fs_ops.c @@ -1576,6 +1576,7 @@ void gfs2_free_block(struct gfs2_sbd *sdp, uint64_t block) rgd->rg.rg_free++; /* adjust the free count */ gfs2_rgrp_out(&rgd->rg, bh->b_data); /* back to the buffer */ brelse(bh, updated); /* release the buffer */ + sdp->blks_alloced--; } /** @@ -1633,6 +1634,9 @@ int gfs2_freedi(struct gfs2_sbd *sdp, uint64_t diblock) /* Set the bitmap type for inode to free space: */ gfs2_set_bitmap(sdp, ip->i_di.di_num.no_addr, GFS2_BLKST_FREE); inode_put(ip, updated); + /* inode_put deallocated the extra block used by the disk inode, */ + /* so adjust it in the superblock struct */ + sdp->blks_alloced--; /* Now we have to adjust the rg freespace count and inode count: */ rgd = gfs2_blk2rgrpd(sdp, diblock); /* The rg itself is in memory as rgd->rg, but there's most likely a */ @@ -1643,5 +1647,6 @@ int gfs2_freedi(struct gfs2_sbd *sdp, uint64_t diblock) rgd->rg.rg_dinodes--; /* one less inode in use */ gfs2_rgrp_out(&rgd->rg, bh->b_data); brelse(bh, updated); /* release the buffer */ + sdp->dinodes_alloced--; return 0; }

1 0

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

cluster-commits February 2010