Re^3: Perl 5.8.8 threads, clean exit

[BrowserUk]

The best suggestion I can come up with given a) your need to stay with 5.8.8; and b) the still scant information of what the threads are doing; is:

1. discard all the signaling stuff; it isn't doing you any good anyway.
2. detach the threads as you create them.
3. Have your main thread sleep for the timeout period and then just fall off the end of the script.

   When the main thread ends; all the others will end also.

   foreach $lpar ( @lparList ) { 
       threads->create( 
           sub { 
               threads->yield(); 
               manageLPAR ($action,$lpar); 
           }
       )->detach; 
   }
   
   sleep $timeout;
   
   ## fall of the end of the script
   ## the detach threads will end when the main thread does.
   

   I rarely advocate abandoning threads this way; but since even if the signals 'worked', you would have no possibility for cleanup as you could never know what your code was doing when the signal to exit arrived; it makes no difference in this case.

[T_I]

This is an option, but then I can better al together abandon threads. The work that's done in this thread has to end before I can execute the fallback option.

This function askes the virtual system to die, the friendly way, via the management server. When it's not dead in the allotted timeframe, the command has to be killed so I can issue the unfriendly command to turn the system off. (i.e. pull the virtual plug) The system (hatrdware the virtual machine is running on) won't accept a second command for the virtual system when there is still a command being processed.

I was looking at threads to solve this issue with a neat single script, to replace the bunch of KSH scripts I now have and that are to slow, but as it seems this is not possible. (with this old perl/threads module) To bad.

[BrowserUk]

I was looking at threads to solve this issue with a neat single script, to replace the bunch of KSH scripts I now have and that are to slow, but as it seems this is not possible. (with this old perl/threads module) To bad.

Actually, it is probably quite easy to solve using even that ancient perl; given enough information about what you are trying to do.

For example:

sub thread {
    ## spawn a process to run the first command
    my $pid = open SSH, '|-', 'ssh' or die $!;

    ## send the command
    print SSH 'the first command';

    ## Sleep for a while
    sleep $timeout;

    ## if the process is still running
    if( kill 0, $pid ) {

        ## kill it
        kill 9, $pid;

        ## wait until it goes away
        sleep 1 while kill 0, $pid;

        ## Open a new ssh session
        $pid = open SSH, '|-', 'ssh' or die $!;
    }

    print SSH 'the second command';
    close SSH;
}

Social

[T_I]

What I want to do is the following (in wierd prototype):

main:

our %ssh (hash of active ssh sessions, Net::OpenSSH)
read file with list of vm's to turn off.
for each line with 1 or more vm's
    start thread1

kill_all_thread1 after timeout

for each line with 1 or more vm's
    start thread2


thread1:
@vms = split (',',$0)
for each vm
    gather all info needed
    if (! exists $ssh{vm-manager}
     { $ssh{vm-manager} = Net::OpenSSH->new(vm-manager) 

       (stdout,stderr)=$ssh{vm-manager}->capture2(request vm to stop)
       *check errors*

       (status,stderr)=$ssh{vm-manager}->capture2(request status vm)
       while (status != dead)
           { (status,stderr)=$ssh{vm-manager}->capture2(request status
+ vm)
             sleep some time }


thread2:
@vms = split (',',$0)
for each vm
    gather all info needed
    if (! exists $ssh{vm-manager}
     { $ssh{vm-manager} = Net::OpenSSH->new(vm-manager) 

       (stdout,stderr)=$ssh{vm-manager}->capture2(pull plug vm)

The current solution with ksh has several problems:

• It's all sequential.
  Not a problem with 2-3 vm's, but the total we have is 70+.
• It initiates a new ssh session per command issued.
  When trying to split the list into seperate commands, the vm-manager gets swamped with ssh sessions. (old hardware, no budget to replace with something fast enough to handle lots of ssh logins)
• Most vms can be stopped parallel, but some combination of vms may only be stopped in sequence. (application requirements)
• Sometimes a vm refuses to die within the allotted time.
• When a vm refuses to die and there are dependent vms, the nice version has to be killed and the rude version has to be started for the same set.
  I.e. in list lpar1,lpar2 when lpar1 refuses to die, it's not allowed to stop lpar2 and then kill lpar1, but it's required to kill lpar1 and then stop/kill lpar2.

Now I'm typing this, I'm thinking of relocating the timeout inside the while loop of thread1, combine thread 1 and 2 and have the resulting thread handle the stopping of it's list. When a timeout within the thread occurs, kill and be friendly with the next. (or surrender after second timeout and exit with a list of vms that have not yet been disabled.)

With the scenario above, is it possible to have a second timeout in main and just simply exit when the main timeout is reached?

$_->join for @jobs;

This waits for the threads to end, how to escape this line? As far as I can find, that's where I need the signal (and the threads->exit() in the signal handler) for.