Patch Name: PHCO_19572

Patch Description: s700_800 10.20 Cumulative vxdump patch

Creation Date: 99/08/18

Post Date:  99/08/20

Hardware Platforms - OS Releases:
	s700: 10.20
	s800: 10.20

Products: N/A

Filesets:
	JournalFS.VXFS-BASE-RUN

Automatic Reboot?: No

Status: General Release

Critical: No

Path Name: /hp-ux_patches/s700_800/10.X/PHCO_19572

Symptoms:
	PHCO_19572:
	vxdump dumps core, hangs or fails with "inode
	bmap failure" messages for large filesystems.

	PHCO_14214:
	A vxdump error will cause the subsequent restore
	to return "Checksum errors".

	PHCO_12186:
	A vxdump process could hang forever.

	PHCO_11556:
	vxdump aborts after one hour if it does not receive a
	response to the following query for mounting the next
	volume:  "vxdump: Is the new volume mounted and ready
	to go?: ("yes" or "no") ".  The following messages are
	displayed before aborting:
	"vxdump: excessive operator query failures"
	"vxdump: The ENTIRE dump is aborted."

	PHCO_11054:
	A vxdump session intermittently hangs forever.
	This was seen when several vxdump sessions were
	started to backup different file systems to different
	tape drives at the same time.  Sometimes one of the
	sessions does not finish and hangs.

Defect Description:
	PHCO_19572:
	Vxdump was incorrectly handling sparse files that
	have gone into multiple indirections. One of the
	levels of indirection must contain more than one entry
	for this problem to occur. The code attempts to track
	the offset of the next indirect extent in intermediate
	levels such that a hole at the end of an indirect can be
	trimmed properly. A incorrect calculation of offset
	resulted in a vxdump core dump and/or "inode bmap
	failure" messages.

	PHCO_14214:
	An incorrect conversion of disk block to file system
	block will result in vxdump generating a corrupt
	dump image. The error will surface during the
	vxrestore of this corrupted dump image. Vxrestore
	will generate checksum errors.

	PHCO_12186:
	vxdump forks a child process which is the master process
	responsible for selecting the files to back up.  This
	master process forks three children (slaves) which take
	turns reading files and then writing them to tape.  The
	signal which is used to communicate between the slaves
	is not handled properly, such that a slave could possibly
	wait forever to receive a signal.  This results in vxdump
	hanging.

	PHCO_11556:
	vxdump aborts after an hour if it does not receive a
	response to the following prompt for next volume:
	"vxdump: Is the new volume mounted and ready to
	go?: ("yes" or "no") ".  The following messages are
	displayed before vxdump aborts:
	"vxdump: excessive operator query failures"
	"vxdump: The ENTIRE dump is aborted."
	The timeout works such that it queries for the next
	volume, and then sends a beep to the console every two
	minutes to alert the operator that it needs attention.
	After 30 times (every 2 minutes = 1 hour), vxdump
	aborts.

	PHCO_11054:
	A vxdump session is comprised of 5 vxdump processes.
	The parent vxdump forks a child vxdump and just waits
	for this child to finish.  This child becomes the master
	vxdump process which is responsible for selecting the files
	to backup.  It creates three child (slave) processes and
	has a separate unnamed pipe through which it communicates
	to each slave.  The slaves work in a round-robin fashion
	by reading the selected file and writing to the tape.
	The slave that has just finished using the tape notifies
	the next slave that it can use the tape by sending it a
	signal.  The problem is that this signal is not always
	caught correctly and all three slaves end up waiting,
	causing the vxdump session to hang forever.

SR:
	4701426866 5003405266 1653215616 1653211300

Patch Files:
	/usr/lbin/fs/vxfs/vxdump
	/usr/sbin/rvxdump
	/usr/sbin/vxdump

what(1) Output:
	/usr/lbin/fs/vxfs/vxdump:
		$Revision: 78.5.1.6 $
		PATCH_10_20: dumprmt.o itime.o main.o optr.o tape.o
			traverse.o unctime.o hpux_rel.o 99/08/18
	/usr/sbin/rvxdump:
		$Revision: 78.5.1.6 $
		PATCH_10_20: dumprmt.o itime.o main.o optr.o tape.o
			traverse.o unctime.o hpux_rel.o 99/08/18
	/usr/sbin/vxdump:
		$Revision: 78.5.1.6 $
		PATCH_10_20: dumprmt.o itime.o main.o optr.o tape.o
			traverse.o unctime.o hpux_rel.o 99/08/18

cksum(1) Output:
	2306897850 98304 /usr/lbin/fs/vxfs/vxdump
	2306897850 98304 /usr/sbin/rvxdump
	2306897850 98304 /usr/sbin/vxdump

Patch Conflicts: None

Patch Dependencies: None

Hardware Dependencies: None

Other Dependencies: None

Supersedes:
	PHCO_11054 PHCO_11556 PHCO_12186 PHCO_14214

Equivalent Patches: None

Patch Package Size: 160 KBytes

Installation Instructions:
	Please review all instructions and the Hewlett-Packard
	SupportLine User Guide or your Hewlett-Packard support terms
	and conditions for precautions, scope of license,
	restrictions, and, limitation of liability and warranties,
	before installing this patch.
	------------------------------------------------------------
	1. Back up your system before installing a patch.

	2. Login as root.

	3. Copy the patch to the /tmp directory.

	4. Move to the /tmp directory and unshar the patch:

		cd /tmp
		sh PHCO_19572

	5a. For a standalone system, run swinstall to install the
	    patch:

		swinstall -x autoreboot=true -x match_target=true \
			-s /tmp/PHCO_19572.depot

	By default swinstall will archive the original software in
	/var/adm/sw/patch/PHCO_19572.  If you do not wish to retain a
	copy of the original software, you can create an empty file
	named /var/adm/sw/patch/PATCH_NOSAVE.

	WARNING: If this file exists when a patch is installed, the
	         patch cannot be deinstalled.  Please be careful
		 when using this feature.

	It is recommended that you move the PHCO_19572.text file to
	/var/adm/sw/patch for future reference.

	To put this patch on a magnetic tape and install from the
	tape drive, use the command:

		dd if=/tmp/PHCO_19572.depot of=/dev/rmt/0m bs=2k

Special Installation Instructions: None