Re: pr00 hanging in "parallel recovery control message reply"

From: Mladen Gogala <gogala.mladen_at_gmail.com>
Date: Thu, 26 Nov 2020 15:01:19 -0500
Message-ID: <670e6ca1-f4d1-facc-acac-c1cb2e2f104f_at_gmail.com>



That's strange. Here is what I have in the man page:

PSTACK(1) Linux Programmer's Manual                 PSTACK(1)

NAME
        pstack - print a stack trace of running processes

SYNOPSIS
        pstack pid [...]

DESCRIPTION
        pstack  attaches  to the active processes named by the pids on the com‐
        mand line, and prints out an execution stack trace, including a hint at
        what  the function arguments are.  If symbols exist in the binary (usu‐
        ally the case unless you have run strip(1)),  then symbolic  addresses
        are printed as well.

        If  the process is part of a thread group, then pstack will print out a
        stack trace for each of the threads in the group.

RESTRICTIONS
_*pstack currently works only on Linux*_, only on an x86 machine running 32         bit  ELF  binaries (64 bit not supported).  Also, for symbolic informa‐

        tion, you need to use a GNU compiler to generate your program, 
and  you
        can't  strip  symbols  from the binaries.  For thread 
information to be
        dumped, you have to use the debug-aware  version  of the  LinuxThreads
        libpthread.so  library.  (To check, run nm(1) on your pthreads library,
        and make sure that the symbol  "__pthread_threads_debug" is  defined.)
        Threads are not supported with the newer NPTL libpthread.so library.

SEE ALSO
        nm(1), ptrace(2)

AUTHORS
        Ross Thompson <ross_at_whatsis.com>

        Red Hat, Inc. <http://bugzilla.redhat.com/bugzilla>

Red Hat Linux                     Feb 25 2002                        PSTACK(1) Apparently, the man page doesn't tell the truth, the whole truth and nothing but the truth, as it is supposed to do. My confidence in the world just took a hard blow. What kind of the world do we live in, when one cannot believe the man pages? BTW, the man page was created in 2002, few weeks ago.

Happy Thanksgiving or what's left of it, with the latest restrictions.

On 11/26/20 2:43 PM, Noveljic Nenad wrote:
>
> It's Solaris x64 11.4.
>
> Support filed the bug: 31716314 : poll/nanosleep waits weeks when
> passed 0.01 second timer values passed.
>
> The fix should be included in the next SRU (SRU27) which should be out
> any minute.
>
> Best regards,
>
> Nenad
>

-- 
Mladen Gogala
Database Consultant
Tel: (347) 321-1217


--
http://www.freelists.org/webpage/oracle-l
Received on Thu Nov 26 2020 - 21:01:19 CET

Original text of this message