Discussion:
DECThreads hard landing
(too old to reply)
Arne Vajhøj
2021-05-11 23:42:56 UTC
Permalink
Maybe first CTRL/C and then CTRL/Y is a bit rough, but still
what happened surprised me.

Arne


$ python recv.py
No handlers could be found for logger "stomp.py"
7

7

7

Cancel

Interrupt

$ dir
%DECthreads bugcheck (version V3.22-095), terminating execution.
% Reason: lckMcsLock: deadlock detected, cell = 0x00000000018B1240
% Running on OpenVMS V8.4-2L2() on AlphaServer DS20E 833 MHz, 1024Mb; 1
CPUs, pid 1097
% The bugcheck occurred at 11-MAY-2021 18:03:25.00, running image
% ARNE1$DKA1:[NET.JFP.python2718.][vms.bin]python.exe;3 in process 00000449
% (named "_TNA3:"), under username "ARNE". AST delivery is enabled for all
% modes; ASTs are active in user. Upcalls are disabled. Multiple kernel
% threads are disabled.
% The current thread sequence number is 2, at 0x018B1240
% Current thread traceback:
% 0: PC 0x80A39694, FP 0x018AE1A0, DESC 0x7BB76B20
% 1: PC 0x80A3FA5C, FP 0x018AE260, DESC 0x7BB77E20
% 2: PC 0x80A3B708, FP 0x018AE310, DESC 0x7BB77458
% 3: PC 0x80A3D5D4, FP 0x018AE330, DESC 0x7BB776C8
% 4: PC 0x80A3BB44, FP 0x018AE360, DESC 0x7BB77730
% 5: PC 0x80A5AFC4, FP 0x018AE3A0, DESC 0x7BB7C2F8
% 6: PC 0x009D9B28, FP 0x018AE3F0, DESC 0x0076E870
% 7: PC 0x009D708C, FP 0x018AE420, DESC 0x0076DC40
% 8: PC 0x800833F8, FP 0x018AE4B0, DESC 0x81951DF8
% 9: PC 0x8017ACF0, FP 0x018AE500, DESC 0x818F21F8
% 10: PC 0x80A34E7C, FP 0x018AE6E0, DESC 0x7BB762B0
% 11: PC 0x80A34C4C, FP 0x018AE740, DESC 0x7BB76540
% 12: PC 0x80A3818C, FP 0x018AE7D0, DESC 0x7BB76640
% 13: PC 0x80A68038, FP 0x018AE800, DESC 0x7BB7DF08
% 14: PC 0x8017ACF0, FP 0x018AE840, DESC 0x818F21F8
% 15: PC 0x801676D4, FP 0x018AEA20, DESC 0x818EFD00
% 16: PC 0x011C1DD4, FP 0x018AEA90, DESC 0x0121DD50
% 17: PC 0x80ABE2C8, FP 0x018AEAE0, DESC 0x7BEA9FE8
% 18: PC 0x00464408, FP 0x018AEB70, DESC 0x00070BC0
% 19: PC 0x0046464C, FP 0x018AEC10, DESC 0x00070C48
% 20: PC 0x00307C24, FP 0x018AEC50, DESC 0x0003D6A0
% 21: PC 0x00346F8C, FP 0x018AECB0, DESC 0x00048A80
% 22: PC 0x00342848, FP 0x018AED20, DESC 0x00048708
% 23: PC 0x00347234, FP 0x018AEE80, DESC 0x00048A80
% 24: PC 0x00342848, FP 0x018AEEF0, DESC 0x00048708
% 25: PC 0x00347234, FP 0x018AF050, DESC 0x00048A80
% 26: PC 0x00342848, FP 0x018AF0C0, DESC 0x00048708
% 27: PC 0x00345894, FP 0x018AF210, DESC 0x00048348
% 28: PC 0x003115CC, FP 0x018AF2E0, DESC 0x0003F390
% 29: PC 0x002F5CA8, FP 0x018AF350, DESC 0x0003A680
% 30: PC 0x0034806C, FP 0x018AF390, DESC 0x00047F40
% 31: PC 0x00342954, FP 0x018AF400, DESC 0x00048708
% 32: PC 0x00347234, FP 0x018AF560, DESC 0x00048A80
% 33: PC 0x00342848, FP 0x018AF5D0, DESC 0x00048708
% 34: PC 0x00347234, FP 0x018AF730, DESC 0x00048A80
% 35: PC 0x00342848, FP 0x018AF7A0, DESC 0x00048708
% 36: PC 0x00345894, FP 0x018AF8F0, DESC 0x00048348
% 37: PC 0x003115CC, FP 0x018AF9C0, DESC 0x0003F390
% 38: PC 0x002F5CA8, FP 0x018AFA30, DESC 0x0003A680
% 39: PC 0x0031A2AC, FP 0x018AFA80, DESC 0x0003FA30
% 40: PC 0x002F5CA8, FP 0x018AFCD0, DESC 0x0003A680
% 41: PC 0x003468EC, FP 0x018AFD10, DESC 0x00047C70
% 42: PC 0x00500B18, FP 0x018AFD50, DESC 0x00089A70
% 43: PC 0x003BB388, FP 0x018AFDA0, DESC 0x0005A890
% 44: PC 0x80A4E74C, FP 0x018AFDE0, DESC 0x7BB7A5A0
% 45: PC 0x80A3B230, FP 0x018AFFE0, DESC 0x7BB772B0
% 46: PC 0x00000000, FP 0x7AE23A50, DESC 0x7BB74380
% 47: PC 0x80340964, FP 0x7AE27B30, DESC 0x81952110
% 48: PC 0x7AF98038, FP 0x7AE27BB0, DESC 0x7AF0D3D0
% Bugcheck output saved to pthread_dump.log.
%SYSTEM-F-IMGDMP, dynamic image dump signal at PC=FFFFFFFF80A39838,
PS=0000001B
Interrupt
Volker Halle
2021-05-12 08:26:48 UTC
Permalink
Arne,

is this what happened ?

You ran your Python program
It seemed to hang, so you issued a CTRL-C
Nothing happened
Then you typed CTRL-Y and got to the DCL prompt
Then you invoked another image ($ DIR), causing the Python image to be terminated forcefully and the DECthreads bugcheck info to be printed and an image dump to be generated - please don't abort the image dump and save the python.dmp file for later analysis.

Is this reproducable ?

Try $ THREADCP /ENABLE=UPCALL python.exe;3 and try to reproduce again.

If you can, please report this problem to VSI support. I can't tell, whether this is an OpenVMS DECthreads problem, a Python problem or a problem with your Python program, but maybe VSI can tell.

Volker.
Arne Vajhøj
2021-05-12 14:26:47 UTC
Permalink
Post by Volker Halle
is this what happened ?
You ran your Python program
It seemed to hang, so you issued a CTRL-C
Nothing happened
Then you typed CTRL-Y and got to the DCL prompt
Then you invoked another image ($ DIR), causing the Python image to be terminated forcefully and the DECthreads bugcheck info to be printed and an image dump to be generated - please don't abort the image dump and save the python.dmp file for later analysis.
Yes.
Post by Volker Halle
Is this reproducable ?
I don't know.

There was a network connectivity problem between the VMS system and the
PC that Python tried to connect to (stomp module trying to connect to
ActiveMQ).

And the Python program was not complete.

I fixed the connectivity issue and completed the program.

It works fine now.

I did not post to get a fix but because it was a rather
surprising outcome.
Post by Volker Halle
If you can, please report this problem to VSI support. I can't tell, whether this is an OpenVMS DECthreads problem, a Python problem or a problem with your Python program, but maybe VSI can tell.
This is way outside my area of expertise.

But I would not expect something like this to happen for a unprivileged
program neither good code nor bad code.

Arne

Volker Halle
2021-05-12 09:32:25 UTC
Permalink
Arne,

there have been reports of DECthreads lckMcsLock deadlocks in the past, which have reportedly been solved by installing PTHREAD patches.

I don't believe there is a PTHREAD patch for VSI OpenVMS Alpha V8.4-2L2, so VSI is your only hope...

Volker.
Loading...