ECO NUMBER: ALPCLUSIO01_062 PRODUCT: OpenVMS Alpha OPERATING SYSTEM 6.2 UPDATE PRODUCT: OpenVMS Alpha OPERATING SYSTEM 6.2 COVER LETTER 1 KIT NAME: ALPCLUSIO01_062 This TIMA kit contains important changes in Volume Shadowing and MOUNT. The new code base is the result of proactive maintenance efforts in these products, and many synchronization and write-logging changes have been implemented. It is also the baseline for all future ECO kits that include an image listed in the "Files Patched" sectio n of this document 2 KITS SUPERSEDED BY THIS KIT: The ALPCLUSIO01_062 remedial kit supersedes the V6.2 Compatibility (COMPAT) save sets that shipped with the V7.1 media. 3 KIT DESCRIPTION: 3.1 Version(s) of OpenVMS to which this kit may be applied: OpenVMS Alpha V6.2, V6.2-1H1, V6.2-1H2, V6.2-1H3 3.2 In order to receive the full fixes listed in this kit the following remedial kits also need to be installed: None 3.3 Files patched or replaced: o [SYSEXE]APB.EXE (new image) o [SYSLIB]CLUE$SDA.EXE (new image) o [SYSEXE]DEBUG_APB.EXE (new image) o [SYSLIB]DISMNTSHR.EXE (new image) o [SYSEXE]DISMOUNT.EXE (new image) o [SYS$LDR]ERRORLOG.EXE (new image) o [SYS$LDR]EXCEPTION.EXE (new image) o [SYS$LDR]EXEC_INIT.EXE (new image) o [SYS$LDR]F11BXQP.EXE (new image) -- COVER LETTER -- Page 2 11 July 1997 o [SYSLIB]INIT$SHR.EXE (new image) o [SYS$LDR]IO_ROUTINES.EXE (new image) o [SYS$LDR]IO_ROUTINES_MON.EXE (new image) o [SYSEXE]LANACP.EXE (new image) o [SYS$LDR]LOCKING.EXE (new image) o [SYSLIB]MOUNTSHR.EXE (new image) o [SYS$LDR]MSCP.EXE (new image) o [SYSLIB]NISCS_LAA.EXE (new image) o [SYSEXE]SDA.EXE (new image) o [SYSEXE]SHADOW_SERVER.EXE (new image) o [SYSEXE]SHADOW_SERVER_DEBUG.EXE (new image) o [SYSEXE]SHOW.EXE (new image) o [SYS$LDR]SYS$CNBTDRIVER.EXE (new image) o [SYS$LDR]SYS$DEBUG_SHDRIVER.EXE (new image) o [SYS$LDR]SYS$DKBTDRIVER.EXE (new image) o [SYS$LDR]SYS$DUDRIVER.EXE (new image) o [SYS$LDR]SYS$ECBTDRIVER.EXE (new image) o [SYS$LDR]SYS$EOBTDRIVER.EXE (new image) o [SYS$LDR]SYS$ERBTDRIVER.EXE (new image) o [SYS$LDR]SYS$ESBTDRIVER.EXE (new image) o [SYS$LDR]SYS$EWBTDRIVER.EXE (new image) o [SYS$LDR]SYS$EXBTDRIVER.EXE (new image) o [SYS$LDR]SYS$EZBTDRIVER.EXE (new image) o [SYS$LDR]SYS$FCBTDRIVER.EXE (new image) o SYS$FRBTDRIVER.EXE (new image) o SYS$FWBTDRIVER.EXE (new image) o [SYS$LDR]SYS$FXBTDRIVER.EXE (new image) -- COVER LETTER -- Page 3 11 July 1997 o [SYS$LDR]SYS$ISLBTDRIVER.EXE (new image) o [SYS$LDR]SYS$KDMBTDRIVER.EXE (new image) o [SYS$LDR]SYS$NISCA_BTDRIVER.EXE (new image) o [SYS$LDR]SYS$PKTBTDRIVER.EXE (new image) o [SYS$LDR]SYS$PKZABTDRIVER.EXE (new image) o [SYS$LDR]SYS$SHDRIVER.EXE (new image) o [SYS$LDR]SYS$TUDRIVER.EXE (new image) o [SYS$LDR]SYS$VM.EXE (new image) o [SYSEXE]SYSBOOT.EXE (new image) o [SYSEXE]SYSINIT.EXE (new image) o [SYS$LDR]SYSLDR_DYN.EXE (new image) o [SYSMSG]SYSMSG.EXE (new image) o [SYS$LDR]SYSTEM_DEBUG.EXE (new image) o [SYS$LDR]TMSCP.EXE (new image) o [SYSEXE]VMOUNT.EXE (new image) o [SYSUPD]VMS$REMEDIAL_ID.EXE (new image) 4 PROBLEMS NOT ADDRESSED IN ALPCLUSIO01_062 KIT o The following problems are not addressed in this kit: o The following three MOUNT problems were discovered at a late stage in the release of this kit. If a user, either manually or by a command procedure, performs one of the following errors, MOUNT may incorrectly add members to existing shadow sets. - A MOUNT/SHAD with an incorrect volume label will succeed in adding the member to the shadow set, for example: $! MOUNT/SYSTEM DSA1/SHAD=$4$DUA1 TST1 $! The shadow set DSA1 is now available with DUA1 as $! the only member $ MOUNT/SYSTEM DSA1/SHAD=$4$DUA5 TST5 $! The device $4$DUA5 is wrongly added as a full copy $! target. -- COVER LETTER -- Page 4 11 July 1997 - Similarly, a MOUNT/SHAD with an incorrect volume label of a shadow set that is mounted elsewhere in the cluster will succeed in adding the member to the shadow set on the other nodes, but the MOUNT will fail on the local node, for example: NODE_1> $ MOUNT/SYSTEM DSA1/SHAD=$4$DUA1 TST1 NODE_1> $ ! The shadow set DSA1 is now available on NODE_1 NODE_2> $ MOUNT/SYSTEM DSA1/SHAD=$4$DUA5 TST5 NODE_2> $! The MOUNT correctly fails on NODE_2 with $! INCVOLLABEL error NODE_1> $! However, the member $4$DUA5 is wrongly added NODE_1> $! to the set DSA1 as a full copy target. - MOUNT will incorrectly allow a non-shareable MOUNT/SHADOW of a disk that is already mounted on another node as "shareable" to succeed. As a result, corruption of the disk(s) will take place, for example: NODE_1> $ MOUNT/SYSTEM DSA1/SHAD=$4$DUA1 TST1 NODE_1> $ ! The shadow set DSA1 is now available on NODE_1 NODE_2> $ MOUNT /NOSHARE DSA5/SHAD=$4$DUA1 TST1 NODE_2> $! The shadow set DSA5 is (wrongly) now available $! on NODE_2 NODE_1> $! The shadow set DSA1 is also available on NODE_1 Corruption of the disk will occur when write operations are performed by either node. OpenVMS Engineering is working on solutions to these problems which will be available in a future MOUNT ECO kit. o RAXCO's PerfectCache V5.0 and PerfectDisk products are not compatible with the MOUNT that is included in CLUSIO and V7.1. If either product is started against a disk, and then another node in the cluster attempts to MOUNT that disk, the MOUNT will fail with a "DEVBUSY" failure. Please note that the SWXNA (the current generation of the FDDI server) presently ships with V5.0 of PerfectCache, and therefore, customers installing this product are exposed to this problem. The HS1xx and HS2xx did not ship with PerfectCache and are not subject to this issue. To workaround this problem, be sure that the disks are mounted on all nodes in the cluster prior to starting PerfectCache or PerfectDisk; or turn off PerfectCache or PerfectDisk prior to mounting the disk on another node. -- COVER LETTER -- Page 5 11 July 1997 RAXCO is aware of this problem and has fixed it in PerfectCache V6. They are working on a solution for PerfectDisk as well. o SHADDETINCON crashes if the database uses a high number of locks (>64K) and Write Logging is in effect on shadow sets. To fix the problem you should install the ALPSHAD06_062 ECO kit. o If one of the shadow set members is not visible at boot time, and the system is booting off the non-master member, the system may crash with a SHADDETINCON bugcheck at SHDRIVER+12124. To work around this problem, the System disk shadow set must be made a single member set prior to booting. After all nodes have booted and are up, the other members can be added to the system disk shadow set. To fix the problem you should install the ALPSHAD06_062 ECO kit. o In a new cluster (never having been booted as a cluster before), if the CNID = 0, and a node leaves the cluster, then all shadow set virtual units in that cluster will go into a full merge. To fix the problem you should install the ALPSHAD06_062 ECO kit. 5 PROBLEMS ADDRESSED IN ALPCLUSIO01_062 A SHADDETINCON BUGCHECK will occur in SHD_THREADS trying to terminate a thread that is still a Significant Event. o The Volume Shadowing driver delivered in V7.1 and the V6.2 Cluster Compatibility kits (xxxCOMPAT_062) did not contain the full solution for the 'Bad Block Repair' (BBR) problem. As a result, when it might be warranted, a disk would not be expelled from the shadow set. o An incompatibility has developed between StorageWorks Host Based RAID Software, and the enhanced volume shadowing provided in both OpenVMS 7.1 and in the Cluster Compatibility Kits (xxxCOMPAT_62). Because of this incompatibility, RAID software can no longer detect that a shadow set state change has occurred. o Write protecting a shadow set member which is being added to an existing shadow set causes virtual unit to hang. o System crashes with INVEXCEPTN Bugchecks in SHSB$SEND_MESSAGE because the UCB address in R5 is zero. Or, it may also crash in IOC_STD$CVT_DEVNAM in IO_ROUTINES when the code tries to get a DDB out of a UCB that is bad. The problem is caused when the IRP$L_ARB field in not correctly set up with the clone error index. Routine SH$VP_DEV_DRVERR uses this byte as an index to fetch the UCB of the erring -- COVER LETTER -- Page 6 11 July 1997 device. The value is FF so an incorrect longword is fetched. The bad value occurs when volume processing initiates mount verification after a device error occurs. o Shadow sets can hang in mountverify for hours after encountering a controller failure (DRAB_INT) on an HSJ50 followed by many 'forced error flagged in last sector read' error messages on multiple shadow set member disks. o Various system crashes in DUDRIVER during connection breaks and reconnections due to PERM CDRP issues. Crash footprints include the following: o INVEXCPTN at DUTU$INIT_MSCP_MSG_C+24 o ACCVIO at SYS$DUDRIVER_NPRO+E274 o ACCVIO at SYS$DUDRIVER_NPRO+E764 o DISKCLASS at DUTU$SEND_MSCP_MSG_49 o INVEXCEPTN at DUTU$INIT_MSCP_MSG_C+24 o INVEXCEPTN at SYS$TUDRIVER_NPRO+109A4 o INVEXCPTN at EXE_STD$PRIMITIVE_FORK_WAIT_C+00070 o INVEXCPTN at SYS$DUDRIVER_NPRO+E0C4 o ADDALRQSZ at EXE$DEANONPGDSIZ+16 o BADDALRQSZ at EXE$DEANONPGDSIZ+12 o INVEXCEPTN at SYS$DUDRIVER_NPRO+0E0A4 o INVEXCEPTN at SYS$DUDRIVER_NPRO+0E324 o The ALPCLUSIO01_062 remedial kit includes the functionality shipped with the ALPCOMPAT_062 Cluster Compatibility Kit. The OpenVMS Cluster Compatibility Kit provides many OpenVMS Version 7.1 enhancements for Version 6.2 systems. This kit is required for Version 6.2 systems if they are included in a cluster with Version 7.1 systems (same system architecture or a mix of VAX and Alpha systems). Optionally, users can install it on other OpenVMS Version 6.2 systems to derive the same benefits. Cluster Compatibility Kit Features: o OpenVMS Version 7.1 Volume Shadowing enhancements The volume shadowing enhancements include significant quality improvements and an increase in supported shadow set members from 400 to 500. Note that the Version 7.1 volume shadowing system disk minimerge feature is not included in the Cluster Compatibility Kit nor is the Dump file off the -- COVER LETTER -- Page 7 11 July 1997 system disk for OpenVMS Alpha. (Dump file off the system disk has been available for OpenVMS VAX systems since Version 6.2.) _______________________ Note _____________________________ If you use volume shadowing, be sure to read the volume shadowing release notes. __________________________________________________________ o OpenVMS Version 7.1 Mount enhancements The Mount utility has been completely rewritten, resulting in a faster, more robust utility. o Correction to an OpenVMS Version 6.2 lock manager problem. The lock manager changes correct a problem in OpenVMS Version 6.2 that could corrupt some internal states in lock information used by fork lock routines, notably the I/O cache subsystem. This problem was corrected in OpenVMS Version 7.0, and also in the limited hardware releases for OpenVMS Alpha Version 6.2-1H2 and Version 6.2-1H3. o Limited support for new port allocation classes for SCSI devices Port allocation classes are a new naming option for SCSI devices on systems running OpenVMS Alpha Version 7.1. If you have installed the Cluster Compatibility Kit (on a VAX or Alpha system ), you can access SCSI disks on an OpenVMS Alpha Version 7.1 system that use port allocation classes in their names, but you cannot name SCSI disks on a Version 6.2 system with port allocation classes. Check with your Customer Support Representative for additional port allocation issues. o System Dump Analyzer Utility (SDA) A special version of the OpenVMS Version 6.2 System Dump Analyzer (SDA) utility is included in the Cluster Compatibility Kit. It recognizes the new volume shadowing data structures. When you install the Cluster Compatibility Kit, the existing OpenVMS Version 6.2 SDA is renamed SDA_OLD.EXE and the Cluster Compatibility Kit version is named SDA.EXE. Use SDA_OLD.EXE to analyze crash dumps from an OpenVMS Version 6.2 system that has not installed the Cluster Compatibility Kit. Use SDA.EXE to analyze crash dumps from an OpenVMS Version 6.2 system that has installed the Cluster Compatibility Kit. -- COVER LETTER -- Page 8 11 July 1997 6 KIT INSTALLATION RATING: The following kit installation rating, based upon current CLD information, is provided to serve as a guide as to which customers should apply this remedial kit. (Reference attached Disclaimer of Warranty and Limitation of Liability Statement) INSTALLATION RATING: INSTALL_2 : To be installed by all customers using the following feature(s): - Customers running the ALPCOMPAT_062 - Customers running OpenVMS V6.2 that want to take advantage of the included enhancements; SHADOW and MOUNT. - Any customer running a VMS Cluster system. This kit is a baseline kit for any OpenVMS Alpha system running in an OpenVMS cluster environment. Future OpenVMS Alpha remedial kits will not install on clustered systems unless the ALPCLUSIO01_062 kit has been installed first. 7 INSTALLATION INSTRUCTIONS: **** NOTE **** Before installing the ALPCLUSIO01_062 remedial kit you should review the following installation notes/warnings. These notes may require you to take action to insure system functionality after installation of the ALPCLUSIO01_062 remedial kit. o The ALPCLUSIO01_062 kit contains all the fixes included in the following list of remedial kits. Once the ALPCLUSIO01_062 kit is installed, you no longer need to install these remedial kits on V6.2* OpenVMS Alpha systems. ALPBOOT05_062 ALPDISM01_062 ALPDRIV04_070 ALPF11X03_070 ALPINIT01_070 ALPMOUN01_062 ALPMSCP01_070 ALPSHAD05_062 ALPSYS02_062 ALPSYS03_062 ALPSYS06_062 ALPSYS08_070 o The ALPCLUSIO01_062 kit contains boot support for the DE500-AA network adaptor. However, to gain full runtime support you must install the ALPLAN04_062 ECO kit. o The Volume Shadowing software which was shipped in OpenVMS VAX and Alpha V7.1 requires additional non-paged pool to improve synchronization. It uses approximately 800 bytes additional non-paged pool per concurrent IO to the virtual unit. Customers should take this into account when they are tuning their systems, and be aware that Volume Shadowing is now more sensitive to resource problems with the possibility that -- COVER LETTER -- Page 9 11 July 1997 systems may crash if non-paged pool is exhausted. o Due to a conflict between the SYSMSG.EXE files in the ALPCLUSIO01_062 ECO Kit and the COBOLRTL024 Kit that has been released by way of the Consolidated Distribution, installation of the ALPCLUSIO01_062 and COBOLRTL024 ECO kits may result in image problems. If the COBOLRTL024 kit has been installed prior to installation of the ALPCLUSIO01_062 kit, the error seen is an image regression error which causes the ALPCLUSIO01_062 installation to abort: 47) SYSMSG (new Image) This kit will NOT be installed. A regression in the level of layered product functionality in [sysmsg]sysmsg.exe would occur as a result of applying an additional level of bugfixes. Existing bugfix/LP level for 1L/0B The kit bugfix/LP level for NE/00 Please contact your Digital ... for a new kit %VMSINSTAL-E-INSFAIL, The installation of ALPCLUSIO01_ V6.2 has failed. This installation failure can be resolved by taking the following steps: - Rename SYS$COMMON:[SYSMSG]SYSMSG.EXE to SYS$COMMON:[SYSMSG]SYSMSG.EXE_HOLD - Install ALPCLUSIO01_062 - *DO NOT* rename the SYSMSG.EXE_HOLD image back to SYSMSG.EXE. The image in the ALPCLUSIO01_062 kit is the most recent image and should be left available to the system. If the COBOLRTL024 kit is installed after the ALPCLUSIO01_062 kit the SYSMSG.EXE in the COBOL kit will overwrite that SYSMSG.EXE from the ALPCLUSIO01_062 kit. In order to preserve the ALPCLUSIO01_062 SYSMSG.EXE image perform the following steps: - Rename SYS$COMMON:[SYSMSG]SYSMSG.EXE to SYS$COMMON:[SYSMSG]SYSMSG.EXE_HOLD - Install COBOLRTL024 - Rename SYS$COMMON:[SYSMSG]SYSMSG.EXE_HOLD to SYS$COMMON:[SYSMSG]SYSMSG.EXE In either case you will lose a COBOL message. In order to restore all messages you should install the ALPDISM02_062 remedial kit. -- COVER LETTER -- Page 10 11 July 1997 Install this kit with the VMSINSTAL utility by logging into the SYSTEM account, and typing the following at the DCL prompt: @SYS$UPDATE:VMSINSTAL ALPCLUSIO01_062 [location of the saveset] The saveset location may be a tape drive, or a disk directory that contains the kit saveset. The system MUST be rebooted after successful installation of the kit and BEFORE any other ECO kit is installed. If the re-boot is not performed, the system will lose functionality. If you have other nodes in your VMScluster, they should also be rebooted in order to make use of the new image(s). Attempts to MOUNT devices before rebooting the system will fail. This includes implicet MOUNTs from BACKUP and VMSINSTAL. If you need to remove the images supplied in this kit, be sure to perform a WRITEBOOT command prior to rebooting your system. Copyright (c) Digital Equipment Corporation, 1997 All Rights Reserved. Unpublished rights reserved under the copyright laws of the United States. The software contained on this media is proprietary to and embodies the confidential technology of Digital Equipment Corporation. Possession, use, or dissemination of the software and media is authorized only pursuant to a valid written license from Digital Equipment Corporation. DISCLAIMER OF WARRANTY AND LIMITATION OF LIABILITY THIS PATCH IS PROVIDED AS IS, WITHOUT WARRANTY OF ANY KIND. ALL EXPRESS OR IMPLIED CONDITIONS, REPRESENTATIONS AND WARRANTIES, INCLUDING ANY IMPLIED WARRANTY OF MERCHANTABILITY, FITNESS FOR PARTICULAR PURPOSE, OR NON-INFRINGEMENT, ARE HEREBY EXCLUDED TO THE EXTENT PERMITTED BY APPLICABLE LAW. IN NO EVENT WILL DIGITAL BE LIABLE FOR ANY LOST REVENUE OR PROFIT, OR FOR SPECIAL, INDIRECT, CONSEQUENTIAL, INCIDENTAL OR PUNITIVE DAMAGES, HOWEVER CAUSED AND REGARDLESS OF THE THEORY OF LIABILITY, WITH RESPECT TO ANY PATCH MADE AVAILABLE HERE OR TO THE USE OF SUCH PATCH.