Conference turris::digital_unix

    That's pretty weird!
    
    Can you grab some kernel stack traces of a sampling of the hung
    ps commands? The only time I've seen ps hang is on VM operations.
    In particular, some unrelated process has a vm map locked for
    write (because of another problem), and ps is the innocent victim
    when it blocks on that locked map.

    Denham,
    
    Thanks for the reply. I'm not familiar with dbx.  I hope the
    following information may help. If not, please post the commands I
    should use.
    
    Regards,
    
    ------------------------------------------------------------------------
    
    (capi105):ttyq3:root:/home/root> ps
      PID TTY      S           TIME COMMAND
      989 console  IW +     0:00.09 /usr/sbin/getty /dev/console console
    vt100
    13451 ttyp1    U  +     0:00.04 ps -t ttyp1     <--- hanged             
    13464 ttyp1    U  +     0:00.04 ps -t ttyp1     <--- hanged
    14307 ttyp1    U  +     0:00.04 ps -t ttyp1     <--- hanged
     8658 ttypc    I  +     0:00.09 /bin/ksh
    13752 ttypd    IW       0:00.36 -ksh (ksh)
    14187 ttypd    U  +     0:00.11 ps -e           <--- hanged
    16192 ttypf    IW       0:00.36 -ksh (ksh)
    16309 ttypf    U  +     0:00.12 ps -e           <--- hanged
    13592 ttyq0    I  +     0:00.23 telnet unix01
    30407 ttyq0    I        0:00.10 /bin/ksh
    15129 ttyq2    IW       0:00.44 -ksh (ksh)
    18390 ttyq2    U  +     0:00.10 ps -e           <--- hanged
    18260 ttyq3    S        0:00.35 -ksh (ksh)
    18389 ttyq3    R  +     0:00.05 ps
    18493 ttyq4    IW       0:00.34 -ksh (ksh)
    18674 ttyq4    U  +     0:00.05 ps a            <--- hanged
     9897 ttyq5    I        0:00.26 /bin/ksh
    12696 ttyq5    S  +     0:00.86 telnet unix06
    13609 ttyq6    I  +     0:00.18 /bin/ksh
    10012 ttyq7    I  +     0:00.08 /bin/ksh
    (capi105):ttyq3:root:/home/root> dbx -k /vmunix
    dbx version 3.11.8
    Type 'help' for help.
    
    stopped at  [thread_block:1906 ,0xfffffc0000470be8]      Source not
    available
    
    warning: Files compiled -g3: parameter values probably wrong
    (dbx) set $pid=18390
    (dbx) where
    >  0 thread_block() ["../../../../src/kernel/kern/sched_prim.c":1903,
    0xfffffc00
    00470be4]
       1 u_anon_dupmcopy(0x11fffe000, 0xfffffc00066a5300, 0x6000, 0x2000,
    0xfffffc00
    0039ee10) ["../../../../src/kernel/vm/u_mape_anon.c":1459,
    0xfffffc0000383000]
       2 u_anon_dup(0x28, 0xffffffff9195f348, 0xfffffc0004ed2b80,
    0xfffffc0003d03c20
    , 0xffffffff00000000) ["../../../../src/kernel/vm/u_mape_anon.c":1199,
    0xfffffc0
    000381dd4]
       3 u_map_copyin(0xfffffc000dc755c0, 0xfffffc000dc75630,
    0xfffffc000dc755c0, 0x
    0, 0xffffffff9195f6d8) ["../../../../src/kernel/vm/vm_umap.c":1949,
    0xfffffc0000
    39f06c]
       4 vm_map_copyin(0xfffffc000dc755c0, 0x0, 0xffffffff9195f6d8,
    0xfffffc0004ed22
    10, 0xfffffc000042321c) ["../../../../src/kernel/vm/vm_map.c":1536,
    0xfffffc0000
    393ac8]
       5 table(0xfffffc0004ed2210, 0x30, 0xffffffff9195f8b8, 0x11fffef08,
    0xfffffc00
    0047c6c8) ["../../../../src/kernel/bsd/cmu_syscalls.c":876,
    0xfffffc0000423218]
       6 syscall(0x1400407f8, 0x1, 0x1, 0x30, 0x55)
    ["../../../../src/kernel/arch/al
    pha/syscall_trap.c":519, 0xfffffc000048c2b4]
       7 _Xsyscall(0x8, 0x3ff800ea7b8, 0x14000ba00, 0x6, 0x2289)
    ["../../../../src/k
    ernel/arch/alpha/locore.s":1094, 0xfffffc000047c4b4]
    (dbx)
    (dbx) p utsname
    struct {
        sysname = "OSF1"
        nodename = "capi105.cgod.telebahia.br"
        release = "V3.2"
        version = "148"
        machine = "alpha"
    } 
    (dbx) quit
    (capi105):ttyq3:root:/home/root> 
    

    Just what we needed. It is indeed a VM problem of some kind.
    Make sure you've got all the VM-related 3.2C patches.

    Denham,
    
    Could you point me the patches I have to apply ?
    I received some time ago tape with 30MB DU patches...
    
    Thanks for your help.
    Regards,

    Well, I was you'd ask me that. I don't know for a fact that
    the problem -- whatever it is -- is fixed. Frankly, I don't have
    the time to sift through the README either. To know what
    to look for, we may have to find the process that's stuck with
    its vm map locked.
    
    Try this:
    
    As root:
    
    1. script
    
    2. kdbx -k /vmunix
    
    3. trace
    
    4. quit
    
    5. exit
    
    6. post the resulting typescript file, or mail it to me:
    
    	[email protected]
    
    Jeff

>    I received some time ago tape with 30MB DU patches...

According to various notices, you have to apply them all, but don't forget to 
back them all out before your next upgrade.

PJDM

I think you need patch-ID OSF350-280.  This patch is superseded by OSF350-290.

--shashi
[Posted by WWW Notes gateway]

    Thanks to all. 
    I'll install patch OSF350-290.
    
    Regards,