sleep(900) ? delayed query... ?

fuzzy1

Howdy,

Trying to set a marker in a MySQL table on our Primary host (WAMP stack) indicating whether or not there is current User Activity -- ON THAT SERVER -- as a signal to remote Secondary -- tablet pc based servers to skip scheduled database synchronization with the Primary.

The idea was to set the following in our SugarCRM apps' index.php
(which is called with pretty much every page in the application)

/////////// SYNC BLOCKING  /////////////////////////
if($currentModule != 'Inspections'){
$block_apro_sync = mysql_query("UPDATE crm.block_sync SET status = 'yes'") or die(mysql_error());  

echo "<div id='block_sync_div' style='border-collapse: collapse;  z-index:100;position: absolute; top: 250px; left: 500px; visibility:visible'>
<iframe src='block_sync.php' name='block_sync_iframe' id='block_sync_iframe' scrolling='no'    frameborder='0' style='width:300px; height:30px; border-width:0px' ></iframe> ";
}
/////////// END SYNC BLOCKING  /////////////////////////

AND in block_sync.php

// block_sync.php
sleep(900); // 15 minutes
$un_block_apro_sync = mysql_query("UPDATE crm.block_sync SET status = 'no'") or die(mysql_error());  

if($un_block_apro_sync){
echo "block_sync.status = NO <br><br>";
}
else{ echo "block_sync.status = YES <br><br>";
}

The effect would be for crm.block_sync.status to toggle between yes AND no between periods of user activity,
with Secondary systems polling the table before each scheduled synchronization.
If block_sync.status=no --- Sync
else -- Don't Sync.

The Trouble is , sleep() fails to perform as I'd hoped when set somewhere upward(90) seconds though I've yet to identify the breakpoint.
Thought about using javascript time_out somehow but I just don't see it.

Please, Any input would be greatly appreciated.

NogDog

Could you be exceeding the max_execution_time of your PHP configuration? Perhaps you'll want to use the [man]set_time_limit/man function.

fuzzy1

Thanks NogDog.
Pretty sure I read somewhere in all of that where setting max_execution_time = 0 eliminates the time limit but tried that with no affect. Also tried commenting out but no change.

As for using set_time_limit(), I also tried
(with max_execution time=0, max_execution time=600, AND max_execution time commented out)

set_time_limit(600) 
sleep(120);
$un_block_apro_sync = mysql_query("UPDATE crm.block_sync SET status = 'no'") or die(mysql_error());  

if($un_block_apro_sync){
echo "block_sync.status = NO <br><br>";
}
else{ echo "block_sync.status = YES <br><br>";
}

and while it's probable that I misunderstand the proper use of set_time_limit() and max_execution time, my read on it is that it both have basically the same effect so... not too surprising is it, that neither one works? Must be something else?? Hmmm? Or maybe another way of thinking about it entirely ???

jmack159

from a comment in the manual:

Note: The set_time_limit() function and the configuration directive max_execution_time only affect the execution time of the script itself. Any time spent on activity that happens outside the execution of the script such as system calls using system(), the sleep() function, database queries, etc. is not included when determining the maximum time that the script has been running.

so that changing the setting of set_time_limit() wouldn't solve your problem.

you should see what sleep() is returning. false on errors, zero on success

hope this helps.

fuzzy1

Thanks jmack 159
I was sure I had read that somewere -- thinking it was in the set_time_limit() manual NOT burried 12 'comments' deep in the sleep() manual.

Still, what I get from both you and NogDog is that the outlined approach is NOT unreasonable at face value??? Or is that just too big a leep?

I'm wondering if I might have better luck with time_sleep_until() which comments in the manual suggest might be viable despite manual Note: "This function is not implemented on Windows platforms"?

NogDog

If this is being run as a web page (as opposed to a command line application), is it possible that the web server is timing out the request? (Although I suppose I'd expect some sort of relevant error message if it were, not just a blank screen. Guess I'm just throwing out random ideas here. 🙂 )

fuzzy1

Thanks NogDog, of course random ideas are ALWAYS welcome.

I think I have worked out a javascript timeout soulution (requires an additional php file but appears to get the job done. Wondering if sleep(900) was just too good to be true. Still curious though, can anyone point to an example where sleep() is known to work with settings >90 seconds???

bradgrafelman

I agree with NogDog about the webserver situation - I would try running this type of script from the CLI and see if it still stops early.