TCPIP V6.0-25 : NFS crashes or hangs system

[joukj]

Hi all,

I can mount an NFS-share located on a Linux (Fedora 40) system using TCPIP V 6.0-25.

Then I do

$ set def nfs_share:[joukj]
$ dir/date *jjj*
....
$ del *.*;*/bef=25-jul-2024:08:00:56
$ del *hhh*.*;*/bef=26-jul-2024:03:55:34
$ dir/date

In 90% of all cases at the last command the system crashes (OpenVMS8.4-L1 AXP) or the command hangs (OpenVMS9.2-2). If the command hangs it cannot be cancelled and even a reboot/shutdown hangs before it finishes.

I cannot try it on my Itanium (I do not want to restart it).


On my Alpha I did not have the problem with TCPIP-V5.7-13. On the X86 machine I never tried before.

[joukj]

Today I crashed also my Itanium, (V8.4-L3 and TCPIP V6.0-25) when doing on a remote NFS-share:
del *.*;*/bef=1-feb-2024

[roberbrooks]

Please post the first page from the output of
ANAL/CRASH SYS$SYSTEM: <or the location of your dump file>
SDA> CLUE CRASH

[debbee.west]

We have already reproduced this issue in the lab and it has been reported to Engineering. But, to guarantee your issue is the same, you could post

$ anal/crash
SDA> clue crash

Thank you!

[joukj]

Attached the crash info from my X86-machine. On AXP/IA64 it is probably simmilar, but presently there are no dump-files written (I removed that in the past because of of low disk-space)



BTW why must the extension of the attached file be .txt and is .list not allowed?

[joukj]

At a hang I found on the console this today: (OpenVMS V9.2-2) :

Code: Select all

Message from user JOUKJ on RUMBA
Client NFS ACP (26000844) - fatal error detected, performing process exit
Signal arguments:

        00000005
        0000000C
        00000004
        FDFDFEB9
        8007A475
        0000001B

[joukj]

I had hopes that this would be fixed in V6.0-26, but it appears not.