[Info-vax] The PWAIT tool.

[Jan-Erik Söderholm]

IanMiller wrote:
> On 1 Oct, 15:54, Jan-Erik Söderholm <jan-erik.soderh... at telia.com>
> wrote:
>> VAXman- @SendSpamHere.ORG wrote:
>>> In article <jW2xm.11964$U5.164... at newsb.telia.net>, =?ISO-8859-1?Q?Jan-Erik_S=F6derholm?= <jan-erik.soderh... at telia.com> writes:
>>>> Hi.
>>>> I have a process that is in "MUTEX" state.
>>>> I found some tool called PWAIT and tried that.
>>>> It runs inside of SDA (after DEFINE PWAIT$SDA to point
>>>> to the PWAIT$SDA.EXE) like this :
>>>> $ ana/sys
>>>> OpenVMS system analyzer
>>>> SDA> pwait SYMBIONT_669
>>>> PWAIT V1.0 (c) 2004, Ian Miller (mil... at encompasserve.org)
>>>> built on VMS V8.2
>>>> reading symbol tables
>>>> %SDA-W-LINKTIMEMISM, link time of SYS$COMMON:[SYS$LDR]ERRORLOG.STB;1
>>>> ( 3-DEC-2004 09:50) does not match link time of image in system
>>>> (18-JUL-2006 14:39)
>>>> Process PID 000261DC name SYMBIONT_669 No. Threads 1
>>>> Thread 0: state MWAIT AST pending KU active (none) blocked (none)
>>> Mutext WAIT state with AST pending (I'm going to assume that Ian's
>>> KU means a special-kernel USER mode AST).
>>>> Process has been waiting for 00:00:00.18
>>>> Process thread resource wait is ENABLED
>>>> Analyzing process locks
>>>>         Process owns no locks
>>>>         waiting for mutex 824A7E00
>>> That's the address of the MUTEX.
>>> You might try to EXAMINE 824A7E00 or MAP 824A7E00.  It might lend some
>>> more insight as to what the mutex is that it is waiting on.
>> SDA> examine 824A7E00
>> FFFFFFFF.824A7E00:  FF67B470.FF67B470   "p´g.p´g."
>> SDA> map 824A7E00
>> %SDA-E-NOTINIMAGE, address (FFFFFFFF.824A7E00) not within a
>>                     system or user image
>> SDA>
>>
>> The process has the following open files :
>>
>> $ pipe sh dev sys$sysdevice/file | sea sys$pipe SYMBIONT_669
>> SYMBIONT_669    000261DC  [VMS$COMMON.SYSMSG]DCPS$MSG.EXE;1
>> SYMBIONT_669    000261DC  [VMS$COMMON.SYSEXE]DCPS$SMB.EXE;1
>> SYMBIONT_669    000261DC  [VMS$COMMON.SYSLIB]SMBSRVSHR.EXE;1
>> SYMBIONT_669    000261DC  [VMS$COMMON.SYSLIB]CMA$OPEN_LIB_SHR.EXE;1
>> SYMBIONT_669    000261DC  [VMS$COMMON.SYSLIB]CMA$LIB_SHR.EXE;1
>> SYMBIONT_669    000261DC  [VMS$COMMON.SYSLIB]DCPS$SMBSHR.EXE;1
>> SYMBIONT_669    000261DC  [VMS$COMMON.SYSLIB]DCPS$BE_SERIAL.EXE;1
>> SYMBIONT_669    000261DC  [VMS$COMMON.SYSLIB]TRN$ANSI_PS.EXE;1
>>
>> so it's an DCPS symbiont, it that helps...
>>
>> When I try SHO CHAN in SDA, SDA hangs. Had to stop/id
>> from another session.
> 
> 
> have you done READ/EXEC, READ SYSDEF, SET PROCESS in SDA to help SDA
> turn the address of the mutex into a symbol?
> AST Pending: KU - means the ast pending longword in the KTB had the
> kernel and user mode bits set.
> 
> PWAIT$SDA - is saying MUTEX wait as the EFWM in the KTB is not
> pointing to the JIB

Hi.

READ/EXEC gave :

SDA> read/exec
%SDA-W-LINKTIMEMISM, link time of SYS$COMMON:[SYS$LDR]ERRORLOG.STB;1
( 3-DEC-2004 09:50) does not match link time of image in system
(18-JUL-2006 14:39)

Actualy the same message that PWAIT gave (as seen above)...

$ HELP /MESS says :

$ h /mess LINKTIMEMISM

  LINKTIMEMISM,  link time of 'file-name' ('dd-mmm-yy hh:mm') does
                 not match link time of image in 'target' ('dd-mmm-yy
                 hh:mm')

   Facility:     SDA, System Dump Analyzer

   Explanation:  The copy of SYS$BASE_IMAGE.EXE used by SDA to build
                 its symbol table is not the same version as that used
                 when the system being analyzed was booted.

                 This message is only a warning. SDA continues unless the
                 mismatch causes an unrecoverable error.

   User Action:  Ensure that logical name ALPHA$LOADABLE_IMAGES points
                 to a directory containing the same version of
                 SYS$BASE_IMAGE.EXE that was used when the system
                 being analyzed was booted.


Right, so :

$ dir SYS$COMMON:[SYS$LDR]ERRORLOG.STB

Directory SYS$COMMON:[SYS$LDR]

ERRORLOG.STB;1   44/54        3-DEC-2004 09:50:59.75

and :

$ sh log  ALPHA$LOADABLE_IMAGES
    "ALPHA$LOADABLE_IMAGES" = "SYS$SYSROOT:[SYS$LDR]" (LNM$SYSTEM_TABLE)

$ dir  ALPHA$LOADABLE_IMAGES:SYS$BASE_IMAGE.EXE

Directory SYS$COMMON:[SYS$LDR]

SYS$BASE_IMAGE.EXE;2      993/1008      6-JAN-2009 15:30:56.84
SYS$BASE_IMAGE.EXE;1      993/1008     15-DEC-2006 07:47:41.45

Total of 2 files, 1986/2016 blocks.

$ write sys$output f$getsyi("boottime")
20-JUL-2009 10:05:06.00


So, is there some trouble here ? Maybe I should delete
SYS$BASE_IMAGE.EXE;1 and rename the ;2 => ;1 ?


READ SYSDEF seems to have worked OK though...

SDA> read sysdef
SDA>

Anyway, these two "reads" and the "set proc" didn't change
the output from the PWAIT command.