Crash dump due to swap file not deinstalling

Post Reply

Topic author
iblowmymind
Contributor
Posts: 17
Joined: Sun Apr 09, 2023 1:54 pm
Reputation: 0
Status: Offline

Crash dump due to swap file not deinstalling

Post by iblowmymind » Sun Apr 09, 2023 2:06 pm

Hello,
I'm faced with an interesting problem using 9.2-1, the messages at the end of shutdown are as follows:

Code: Select all

Open user files prevent dismounting host device _BLWX86$DKA0:
Cannot deinstall page/swap file DISK$X86SYS:[SYS0.SYSEXE]PAGEFILE.SYS;1
         (index # 254)


VSI Dump Kernel SYSBOOT Jan 23 2023 14:03:45
** Dumping error logs to the system disk (BLWX86$DKA0:)
** Error logs dumped to BLWX86$DKA0:[SYS0.SYSEXE]SYS$ERRLOG.DMP
** (used 84 out of 96 available blocks)
This error does not affect regular shutdown/powerup, but I'd still like to figure out why this is happening. I fear it could lead to disk corruption due to the disk not dismounting properly at each shutdown.

I've tried out ELV (ANALYZE /ERR /ELV), which fails to read ERRLOG.DMP with:

Code: Select all

ELV> translate sys$system:sys$errlog.dmp
Output file SYS$OUTPUT: created at  9-APR-2023 21:00:52.32


Output for SYS$SYSROOT:[SYSEXE]SYS$ERRLOG.DMP;2
%ELV-E-INVERLEVT, invalid error log event in SYS$SYSROOT:[SYSEXE]SYS$ERRLOG.DMP;2
-ELV-W-NODNOTFND, bit-to-text node not found
%ELV-I-SKPERLEVT, skipping error log event 1
The error occurs after operator.log is closed, so nothing on there.
Would anyone be able to help?

User avatar

volkerhalle
Master
Posts: 196
Joined: Fri Aug 14, 2020 11:31 am
Reputation: 0
Status: Offline

Re: Crash dump due to swap file not deinstalling

Post by volkerhalle » Mon Apr 10, 2023 1:00 am

Hi,

this is not a crash dump !

SYS$SYSTEM:SYS$ERRLOG.DMP is just a dump of the contents of the current Error Log buffers from memory at the time of system shutdown (or system crash).This file will be read during system boot to preserve errlog messages, which were reported shortly before the shutdown/crash and could not be written to ERRLOG.SYS anymore. Entries from SYS$ERRLOG.DMP are automatically being added to ERRLOG.SYS during startup.

This file cannot be read by ELV. SDA reports (this is an example from an OpenVMS Alpha system):

AXPVMS $ anal/cras sys$system:sys$errlog.dmp

OpenVMS system dump analyzer
%SDA-E-NODUMPHERE, only the error log buffers were written to this file

There is no risk of disk corruption, if the system disk cannot be cleanly dismounted during shutdown. You might only loose some free space on the system disk, which can be recovered by SET VOLUME/REBUILD (or ANALYZE/DISK/REPAIR).

I would be more concerned about the consistency of your user files, if they have not been correctly closed during shutdown.

There should be no open 'user files' at the end of SYSHUTDWN.COM. Maybe add a SHOW SYSTEM and/or SHOW DEV/FILES SYS$SYSDEVICE at the end of SYSHUTDWN.COM and - if there are any open user files - add the appropriate shutdown commands into SYSHUTDWN.COM to stop your applications.

Volker.
Last edited by volkerhalle on Mon Apr 10, 2023 3:32 am, edited 2 times in total.


Topic author
iblowmymind
Contributor
Posts: 17
Joined: Sun Apr 09, 2023 1:54 pm
Reputation: 0
Status: Offline

Re: Crash dump due to swap file not deinstalling

Post by iblowmymind » Wed Apr 12, 2023 3:11 pm

Hi,
Thanks for your response! Yeah, I realised that the dumping logs thing was a normal shutdown message, and not, in fact, a kernel dump. I tried the methods you suggested, but I wasn't able to see any files that weren't supposed to be open at that stage. I only see TCPIP and VMS binaries, and TCPIP gets unloaded after SYSHUTDWN. Pretty sure I can't get to the bottom of it that easily, as console OPA0 gets disabled before the error occurs. Any suggestions?

User avatar

volkerhalle
Master
Posts: 196
Joined: Fri Aug 14, 2020 11:31 am
Reputation: 0
Status: Offline

Re: Crash dump due to swap file not deinstalling

Post by volkerhalle » Sat Apr 15, 2023 4:26 am

Hi,

after upgrading my VSI OpenVMS x86-64 V9.2 system (VMware Workstation Player 16) to E9.2-1, I can confirm your observation. These 2 messages are being issued from OPCCRASH and are NEW in E9.2-1. I've temporarily added a SHOW DEV/FILES SYS$SYSDEVICE and SHO SYS command immediately preceeding the RUN SYS$SYSTEM:OPCCRASH command in SHUTDOWN.COM:

Code: Select all

Files accessed on device DSA0: on 15-APR-2023 09:14:42.74

Process name      PID     File name
                00000000  [000000]INDEXF.SYS;1
                00000000  [VMS$COMMON.SYS$LDR]SYS$MD.DSK;2
                00000000  [VMS$COMMON.SYSLIB]DCLTABLES.EXE;147
SHADOW_SERVER   00000405  [VMS$COMMON.SYSEXE]SHADOW_SERVER.EXE;1
                00000000  [SYS0.SYSEXE]PAGEFILE.SYS;1
LANACP          00000406  [VMS$COMMON.SYSEXE]LANACP.EXE;1
LANACP          00000406  [SYS0.SYSMGR]LAN$ACP.LOG;25
FASTPATH_SERVER 00000408  [VMS$COMMON.SYSEXE]FASTPATH_SERVER.EXE;1
IPCACP          00000409  [VMS$COMMON.SYSEXE]IPCACP.EXE;1
OPCOM           0000040C  [VMS$COMMON.SYSEXE]OPCOM.EXE;1
JOB_CONTROL     0000040E  [VMS$COMMON.SYSEXE]JBC$JOB_CONTROL.EXE;1
JOB_CONTROL     0000040E  [SYS0.SYSMGR]ACCOUNTNG.DAT;1
                00000000  [VMS$COMMON.SYSEXE]TERMTABLE.EXE;1
                00000000  [VMS$COMMON.SYSMSG]TCPIP$MSG.EXE;1
                00000000  [VMS$COMMON.SYSLIB]TCPIP$ACCESS_SHR.EXE;1
TCPIP$INETACP   00000418  [VMS$COMMON.SYSEXE]TCPIP$INETACP.EXE;1
SMHANDLER       0000041C  [SYS0.SYSMGR]SYS$SMHANDLER.LOG;24
SMHANDLER       0000041C  [VMS$COMMON.SYSEXE]SYS$SMHANDLER.COM;1
SMHANDLER       0000041C  [VMS$COMMON.SYSEXE]SYS$SMHANDLER.EXE;1
SYSTEM          0000041D  [VMS$COMMON.SYSEXE]SHUTDOWN.COM;2
SYSTEM          0000041D  [VMS$COMMON.SYSEXE]SHOW.EXE;1
OpenVMS E9.2-1  on node VOLKER   15-APR-2023 09:14:42.80   Uptime  0 00:02:57
  Pid    Process Name    State  Pri      I/O       CPU       Page flts  Pages
00000401 SWAPPER         HIB     16        0   0 00:00:00.18         0     17
00000405 SHADOW_SERVER   HIB      6        7   0 00:00:00.01       157    206
00000406 LANACP          HIB     14       62   0 00:00:00.02       204    255
00000408 FASTPATH_SERVER HIB     10        8   0 00:00:00.01       161    202
00000409 IPCACP          HIB     10        8   0 00:00:00.00       138    180
0000040C OPCOM           HIB      8      108   0 00:00:00.03       157    205
0000040E JOB_CONTROL     HIB      9       47   0 00:00:00.02       181    247
00000418 TCPIP$INETACP   HIB      8      154   0 00:00:00.07       251    232
0000041C SMHANDLER       HIB      8       45   0 00:00:00.00       270    294
0000041D SYSTEM          CUR   0  4     1695   0 00:00:00.55      2385    360

Open user files prevent dismounting host device _DSA0:
Cannot deinstall page/swap file DISK$X86SYS_E921:[SYS0.SYSEXE]PAGEFILE.SYS;1
         (index # 254)

These 2 new messages are at least 'disturbing' to the normal system manager. And one could discuss, if those open files are really 'open user files'.

EDIT 17-APR-2023: Slight correction: these 2 messages already existed in E9.2. I have raised an issue to VSI.

Volker.
Last edited by volkerhalle on Mon Apr 17, 2023 8:21 am, edited 1 time in total.

User avatar

volkerhalle
Master
Posts: 196
Joined: Fri Aug 14, 2020 11:31 am
Reputation: 0
Status: Offline

Re: Crash dump due to swap file not deinstalling

Post by volkerhalle » Wed Apr 19, 2023 8:51 am

Answer from VSI: these 2 messages at the end of shutdown are no reason for concern:

Open user files prevent dismounting host device _BLWX86$DKA0:
Cannot deinstall page/swap file DISK$X86SYS:[SYS0.SYSEXE]PAGEFILE.SYS;1
(index # 254)

They will either be removed or eventually mentioned in the release notes.

Volker.

Post Reply