The 83% kernel time is certainly a clue. I might begin to suspect some problem
with logicals or symbols. I second Dan's call for a list of logicals and
symbols on a system where it works and on a system where it
doesn't work.
-----Original Message-----
From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED]]
Sent: Friday, June 30, 2000 12:47 PM
To: [EMAIL PROTECTED]; [EMAIL PROTECTED];
[EMAIL PROTECTED]; [EMAIL PROTECTED]; [EMAIL PROTECTED]
Subject: RE: Perl hangs on OpenVMS
SHOW PROC/CONT/ID
It jumps around quite a lot.
SHOW PROC/CONT/ID + v
Solid almost all of the time
MON MODES
With Perl running, 83% kernel mode, 16% user mode
Without Perl running, 11% user mode, 86% idle time
SDA> SHOW PROC/ID=
Process index: 00FA Name: S95697 Extended PID: 2B8024FA
-----------------------------------------------------------
Process status: 02040001 RES,PHDRES
Required capabilities: 0000000C QUORUM,RUN
PCB address 81982200 JIB address
819B7B40
PHD address 8D586000 Swapfile disk address
00000000
Master internal PID 001200FA Subprocess count
0
Internal PID 001200FA Creator internal PID
00000000
Extended PID 2B8024FA Creator extended PID
00000000
State COM Termination mailbox
0000
Previous CPU Id 00000000 Current CPU Id
00000000
Previous ASNSEQ 000000000000BB9E Previous ASN
000000000000007E
Current priority 4 # of threads
0000000000000000
Initial process priority 4 Delete pending count 0
Base priority 4 AST's active
NONE
UIC [01004,000000] AST's remaining
4998
Mutex count 0 Buffered I/O count/limit
5000/5000
Waiting EF cluster 0 Direct I/O count/limit
5000/5000
Abs time of last event 04B67CC1 BUFIO byte count/limit
98208/98208
Event flag wait mask 0000000D # open files allowed left
98
Process index: 00FA Name: S95697 Extended PID: 2B8024FA
-----------------------------------------------------------
Swapped copy of LEFC0 C73D00A6 Timer entries allowed left
100
Swapped copy of LEFC1 E8000001 Active page table count
14
Global cluster 2 pointer 00000000 Process WS page count
39
Global cluster 3 pointer 00000000 Global WS page count
2
SDA> SHO PROC/ID=<process ID of hanging Perl>/CHAN
Channel Window Status Device/file accessed
------- ------ ------ --------------------
0010 00000000 DKA104:
0020 81332E80
DKA104:[PERL5_005_03]PERL.EXE;1
0030 8142C3C0
DKA103:[SYSCOMMON.SYSLIB]DECC$SHR.EXE;2 (section file)
0040 8142BBC0
DKA103:[SYSCOMMON.SYSLIB]DPML$SHR.EXE;1 (section file)
0050 814153C0
DKA103:[SYSCOMMON.SYSEXE]DCL.EXE;1 (section file)
0060 81422740
DKA103:[SYSCOMMON.SYSLIB]DCLTABLES.EXE;186 (section file)
0070 8142B540
DKA103:[SYSCOMMON.SYSLIB]CMA$TIS_SHR.EXE;1 (section file)
0080 00000000 FTA82:
0090 00000000 FTA82:
00A0 8142A980
DKA103:[SYSCOMMON.SYSLIB]LIBRTL.EXE;3 (section file)
00B0 8142AB80
DKA103:[SYSCOMMON.SYSLIB]LIBOTS.EXE;1 (section file)
00C0 81303100
DKA104:[PERL5_005_03]PERLSHR.EXE;1
00D0 81498740
DKA103:[SYSCOMMON.SYSLIB]SECURESHRP.EXE;1 (section file)
00E0 814B0B40
DKA103:[SYSCOMMON.SYSLIB]SECURESHR.EXE;1 (section file)
00F0 00000000 FTA82:
> -----Original Message-----
> From: Jordan Henderson [SMTP:[EMAIL PROTECTED]]
> Sent: Friday, June 30, 2000 6:30 PM
> To: '[EMAIL PROTECTED]'; Jordan Henderson;
> [EMAIL PROTECTED]; [EMAIL PROTECTED]; [EMAIL PROTECTED]
> Subject: RE: Perl hangs on OpenVMS
>
> Are the PC addresses when you run a "SHO PROC/CONT/ID=<process ID>"
> hanging in a very tight loop, or does it jump around quite a bit.
>
> Also, when you are in a "SHO PROC/CONT", hit the "v" key. Do you see
> the PC (as represented by the cursor) move around a lot and disappear
> or is it solid in one or two locations.
>
> It might be interesting to capture the output of a a "MON MODES" command
> over a period of a few minutes when this is not running and when it is.
>
> If you have privileges, you might provide information from the following:
>
> >ANAL/SYS
> SDA> SHO PROC/ID=<process ID of hanging Perl>
> SDA> SHO PROC/ID=<process ID of hanging Perl>/CHAN
>
> We're probing around in the dark here. As nobody else has seen this
> behavior, I would look for the possibility of changes on your system.
>
> -----Original Message-----
> From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED]]
> Sent: Friday, June 30, 2000 11:49 AM
> To: [EMAIL PROTECTED]; [EMAIL PROTECTED];
> [EMAIL PROTECTED]; [EMAIL PROTECTED]; [EMAIL PROTECTED]
> Subject: RE: Perl hangs on OpenVMS
>
>
> My Perl is not built to pre-load the ENV array.
>
> The logicals that are set up are defined in a startup procedure. The same
> procedure is called from both the functioning and the non functioning node
> in my cluster.
>
> The Perl -V command works exellent on the "functioning" node, but hangs on
> the "nonfunctioning" node. The distribution was built and has worked for
> several
> months on the now "nonfunctioning" node.
>
> In this case "nonfunctioning" means that Perl has stopped functioning on
> the
> node. Ordinary activities and program development in other languages work
> just as usual.
>
> - AW -
>
> > -----Original Message-----
> > From: Jordan Henderson [SMTP:[EMAIL PROTECTED]]
> > Sent: Friday, June 30, 2000 5:43 PM
> > To: 'BAZLEY, Sebastian'; 'Dan Sugalski'; '[EMAIL PROTECTED]';
> > [EMAIL PROTECTED]
> > Subject: RE: Perl hangs on OpenVMS
> >
> > If it's chasing poorly defined logicals around, then you might
> > see a dramatic increase in "Log Name Translation Rate" when
> > you do a "MON IO" command when this hanging Perl is running.
> >
