Professional Web Applications Themes

Going from multiuser to single-user back to multiuser - SCO

Yesterday I had occasion to bring our 5.0.7 system down to single-user mode using shutdown -g0 su . When done with my maintenance, I exited single-user mode and tried to go back to multi-user mode by entering '2' at the init level (0-6): prompt. A couple of the rc2 startup scripts failed to execute properly: From S35dlpi.log: dlpid: Unable to open network adapter driver (/dev/mdi/eeE0) dlpid: No such file or directory dlpid: DLPI Interface (net0) not started From S95docview.log: Ouch! ap_mm_create(520168, "/usr/lib/docview/logs/httpd.mm.51") failed Error: MM: mm:core: failed to acquire semaphore (No space left on device): OS: Invalid argument /etc/rc2.d/S95docview start: ...

  1. #1

    Default Going from multiuser to single-user back to multiuser

    Yesterday I had occasion to bring our 5.0.7 system down to single-user
    mode using shutdown -g0 su .

    When done with my maintenance, I exited single-user mode and
    tried to go back to multi-user mode by entering '2' at the init level (0-6):
    prompt.

    A couple of the rc2 startup scripts failed to execute properly:

    From S35dlpi.log:

    dlpid: Unable to open network adapter driver (/dev/mdi/eeE0)
    dlpid: No such file or directory
    dlpid: DLPI Interface (net0) not started

    From S95docview.log:

    Ouch! ap_mm_create(520168, "/usr/lib/docview/logs/httpd.mm.51") failed
    Error: MM: mm:core: failed to acquire semaphore (No space left on device):
    OS: Invalid argument
    /etc/rc2.d/S95docview start: httpd could not be started

    Both of these scripts work just fine from a cold boot, whether
    or not you go to single-user mode first.

    rc2.d/S35dlpi (aka rc0.d/K97dlpi) simply runs /etc/lli start or stop,
    depending
    on whether you're entering or leaving multiuser mode. (/etc/lli is identical
    to /etc/nd)

    rc2.d/S95docview does not seem to have a corresponding script
    in rc0.d to shut itself down, probably because the P90apache
    shutdown script accomplishes the same result?

    Is all this the expected system behavior?

    Note that the directory /usr/lib/docview/logs contains:

    -rw------- 1 root root 4 Aug 18 17:33 docview.pid
    -rw------- 1 nouser root 1048576 Jun 22 15:07 httpd.mm.659.mem
    -rw------- 1 nouser root 0 Jun 22 15:07 httpd.mm.659.sem
    -rw------- 1 nouser root 1048576 Jul 7 17:25 httpd.mm.667.mem
    -rw------- 1 nouser root 0 Jul 7 17:24 httpd.mm.667.sem
    -rw------- 1 nouser root 1048576 Jun 22 15:47 httpd.mm.685.mem
    -rw------- 1 nouser root 0 Jun 22 15:47 httpd.mm.685.sem
    -rw------- 1 nouser root 1048576 Jun 22 17:07 httpd.mm.699.mem
    -rw------- 1 nouser root 0 Jun 22 17:07 httpd.mm.699.sem
    -rw------- 1 nouser root 1048576 Jul 25 17:45 httpd.mm.734.mem
    -rw------- 1 nouser root 0 Jul 25 17:45 httpd.mm.734.sem
    -rw------- 1 nouser root 1048576 Jun 22 15:36 httpd.mm.769.mem
    -rw------- 1 nouser root 0 Jun 22 15:36 httpd.mm.769.sem
    -rw-r--r-- 1 root root 0 Jul 31 15:53 navlist_log

    Are the *.[ms]em files simply the result of improper shutdowns, and
    can they be safely deleted?

    Bob


    Bob Bailin Guest

  2. #2

    Default Re: Going from multiuser to single-user back to multiuser

    Bob Bailin typed (on Tue, Aug 19, 2003 at 07:03:23PM +0000):
    |
    | >From S95docview.log:
    |
    | Ouch! ap_mm_create(520168, "/usr/lib/docview/logs/httpd.mm.51") failed
    | Error: MM: mm:core: failed to acquire semaphore (No space left on device):
    | OS: Invalid argument
    | /etc/rc2.d/S95docview start: httpd could not be started

    Boo-boo in MP1.
    Raise the kernel parameter SEMMNI from the default of 10.
    25 will do.

    | Note that the directory /usr/lib/docview/logs contains:
    |
    | -rw------- 1 root root 4 Aug 18 17:33 docview.pid
    | -rw------- 1 nouser root 1048576 Jun 22 15:07 httpd.mm.659.mem
    | -rw------- 1 nouser root 0 Jun 22 15:07 httpd.mm.659.sem
    | -rw------- 1 nouser root 1048576 Jul 7 17:25 httpd.mm.667.mem
    | -rw------- 1 nouser root 0 Jul 7 17:24 httpd.mm.667.sem
    | -rw------- 1 nouser root 1048576 Jun 22 15:47 httpd.mm.685.mem
    | -rw------- 1 nouser root 0 Jun 22 15:47 httpd.mm.685.sem
    | -rw------- 1 nouser root 1048576 Jun 22 17:07 httpd.mm.699.mem
    | -rw------- 1 nouser root 0 Jun 22 17:07 httpd.mm.699.sem
    | -rw------- 1 nouser root 1048576 Jul 25 17:45 httpd.mm.734.mem
    | -rw------- 1 nouser root 0 Jul 25 17:45 httpd.mm.734.sem
    | -rw------- 1 nouser root 1048576 Jun 22 15:36 httpd.mm.769.mem
    | -rw------- 1 nouser root 0 Jun 22 15:36 httpd.mm.769.sem
    | -rw-r--r-- 1 root root 0 Jul 31 15:53 navlist_log
    |
    | Are the *.[ms]em files simply the result of improper shutdowns, and
    | can they be safely deleted?

    Yup, zap 'em.

    --
    JP
    Jean-Pierre Radley Guest

  3. #3

    Default Re: Going from multiuser to single-user back to multiuser

    Jean-Pierre Radley wrote:
    > Bela Lubkin typed (on Thu, Aug 21, 2003 at 01:31:24AM +0000):
    > | Bob Bailin wrote:
    > |
    > | > When done with my maintenance, I exited single-user mode and tried
    > | > to go back to multi-user mode by entering '2' at the init level
    > | > (0-6): prompt.
    > |
    > | That's a prompt that I'm pretty sure you're not really ever supposed
    > | to be able to reach.
    >
    > I've seen it more than once, and fairly recently, when exiting the shell
    > after running 'shutdown -i1'.
    Hmmm. Ok, I guess that's because I hardly ever retreat to single-user
    from multiuser mode, and if I do, I use `init 2` explicitly to get back.
    >Bela<
    Bela Lubkin Guest

  4. #4

    Default Re: Going from multiuser to single-user back to multiuser


    "Bela Lubkin" <com> wrote in message
    news:com... [/ref]

    'su' was always more intuitive than '-i1', and 1 fewer keystroke.
    I never used 'init 1' out of habit and politeness to the users.
     

    shutdown -i1
    gives the exact same results as
    shutdown su

    In both cases, I end up at:
    new run level: S
    and the sulogin prompt

    Running ps -ef in single-user mode shows the same processes
    running in either case, and the only mounted filesystems are
    root and boot.
     [/ref]
    (0-6): 
    >
    > That's a prompt that I'm pretty sure you're not really ever supposed to
    > be able to reach.[/ref]

    As JPR pointed out, you *always* end up at this prompt when you
    exit from single-user mode with ctrl-D, rather than running init 2.
     
    >
    > I'm specifically curious about this one (when using init 1/init 2).
    > [/ref]
    device): 
    >
    > This is due to the default semaphore parameters being inadequate. It's
    > something we should have fixed in OSR507MP1, but we missed it. Change
    > kernel parameter "SEMMNI" to a larger value. The default is 10; 100 is
    > a perfectly reasonable value to raise it to. (The cost is something
    > like 3K of RAM...)
    > [/ref]
    identical 
    >
    > That would be my guess, but I don't really know.

    >
    > The apache/docview failure is "expected" in the sense that it's a known
    > problem with a known fix. The dlpi problem is new to me.[/ref]

    After reading the recent TA about the number of semaphores needed
    after applying 5.0.7 MP1, I decided to take a look at the output of:

    ipcs -a

    both in multiuser and single-user mode. The TA is right, there are a *lot*
    of semaphore structures in use in multi-user mode. The problem seems
    to be that the semaphores are not released when their associated
    programs (docview, etc) are terminated when transitioning to init 1.
    When I manually remove all memory structures with ipcrm while in
    single-user mode, there are no longer any problems starting up
    S95docview when transitioning back to init level 2.

    (All 3 SEM* parameters were already set to 100.)

    However the problem starting S35dlpi remains, except that now,
    only the 1st two lines of the error log listed above are present.

    I'm certain these problems are easily repeatable on any standard
    installation of 5.0.7. (I didn't include the output of ipcs -a because
    the width of the report makes for terrible reading when wrapped
    by your favorite newsreader. I piped my output to a laser printer
    in landscape mode for easy reading.)

    Finally, as a check, I rebooted the system and went directly into
    single-user mode. As expected, ipcs -a showed that no memory
    structures were defined.
     [/ref]
    [snip] 
    >
    > They can be safely deleted (of course, don't delete any that belong to
    > the currently running apache...)[/ref]

    Deleting them in single-user mode avoids any possible problems.
     [/ref]

    Just a reminder, I fully realize that all of the above falls into the
    category
    of "interesting behavior". I doubt it would pose a problem to anyone in
    a real-world situation (i.e., you can always reboot).

    Bob


    Bob Guest

Similar Threads

  1. shockwave3D multiuser test...
    By hondo3000 in forum Macromedia Director 3D
    Replies: 7
    Last Post: January 20th, 01:37 PM
  2. multiuser multimedia environment
    By hilmar hermens in forum Macromedia Director Basics
    Replies: 0
    Last Post: February 22nd, 08:44 PM
  3. Xtra MultiUser
    By :. Buzzalino \(Foros Macromedia\) .: in forum Macromedia Director Lingo
    Replies: 0
    Last Post: November 4th, 07:07 PM
  4. Multiuser
    By Jota Mix in forum Macromedia Director Lingo
    Replies: 0
    Last Post: September 9th, 10:01 PM
  5. Communicating without using multiuser server
    By Word of Mouth Productions in forum Macromedia Director Lingo
    Replies: 3
    Last Post: August 19th, 05:54 PM

Bookmarks

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •  

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139