We have a client server architecture with Angular on client side and Apache2 PHP PDO and MySQL on the server side. server side exposing an API to clients that gives them data to show.
Some observations :
some API calls can take very long to compute and return response.
server side seem to handle a single request per client at any given time (im seeing only one coresponding query thats being executed in mysql), that limit comes either from apache or from mysql since front-end sending requests in parallel for sure.
front end cancels requests that are not relevant anymore (data being fetched will not be visible)
seems like requests canceled by front end are not canceled in server side and continues to run anyway, i think even if they are queued they will still run when their turn arrives (even though they were cancelled on client side)
Need help to understand :
what exactly is the cause of not having all requests (or at least X>1 requests) run on parallel? can it be changed?
What configurations should i change in either apache or mysql to overcome this?
is there a way to make apache drop cancelled requests? at least those that are still queued and not started?
Thanks!
EDIT
Following #Markus AO comment (Thanks Markus!!!) this was session blocking related... wish i knew about that before!
OP has a number of tangled problems on the table. However I feel these are worthwhile concerns (having wrestled with them myself), so let's take this apart. For great justice; main screen turn on:
Solving Concurrent Request Problems
There are several possible problems and solutions with concurrent connections in a (L)AMP stack. Before looking at tuning Apache and MySQL, however, let me gloss a common "mystery" issue that creates concurrence problems; namely, a necessary evil called "PHP Session Locking".
PHP Session Blocking & Concurrent Requests
In a nutshell: When you use sessions in your application, after calling session_start(), PHP locks the session file stored at your session.save_path directory. This file lock will remain in place until the script ends, or session_write_close() is called. Result: Any subsequent calls by the same user will be queued, rather than concurrently processed, to ensure there's no session data corruption. (Imagine parallel scripts writing into the same $_SESSION!)
An easy way to demonstrate this is to create a long-running script; then call it in your browser; and then open a new tab, and call it again (or in fact, call any script sharing the same session cookie/ID). You'll see that the second call won't execute until the first one is concluded. This is a common cause of strange AJAX lags, especially with parallel AJAX requests from a single page. Processing will be consecutive instead of concurrent. Then, 10 calls at 0.3 sec each will take a total of 3 sec to conclude, and so on. We don't want that, do we!
You can remedy request blocking caused by PHP session lock by ensuring that:
Scripts using sessions should call session_write_close() once done storing session data. The session lock will be immediately released.
Scripts that don't require sessions shouldn't start sessions to begin with.
Scripts that need to only read session data: Using session_start() with ['read_and_close' => true] option will give you a read-only (non-persistent) $_SESSION variable without session locking. (Available since PHP 7.)
Options 1 and 3 will leave you with read access for the $_SESSION variable and release/avoid the session lock. Any changes made to $_SESSION after the session is closed will be silently discarded; no warnings/errors are displayed.
The session lock request blocking issue is only consequential for a single user (using the same session). It has no impact on multi-user concurrence. For further reading, please see:
SO: Session (Auto)-Start, Performance & Session Locking
SO: PHP & Sessions: Is there any way to disable PHP session locking?
In-Depth: PHP Session Locking: How To Prevent Sessions Blocking in PHP requests.
Apache & MySQL Concurrent Requests
Once upon a time, before realizing PHP was the culprit behind blocking/queuing my concurrent calls, I spent a small aeon in tweaking Apache and MySQL and wondering, what happen?
Apache 2.4 supports 150 concurrent requests by default; any further requests will queue up. There are several settings under the MPM/Multi-Processing Module that you can tune to support the desired level of concurrent connections. Please see:
MPM Docs
Worker Docs
Overview at Oxpedia
MySQL has options for max_connections (default 151) and max_user_connections (default unlimited). If your application sends a lot of concurrent requests per user, you'll want to ensure the global max connections is high enough to ensure a handful of users don't hog the entire DBMS.
Obviously, you'll further want to tune these settings in light of your server CPU/RAM specs. (The calculations for which are beyond this answer.) Your concurrency issues probably aren't caused by too many open TCP sockets, but hey, you never know...
Canceling Requests to Apache/PHP/MySQL
We don't have much to go on as far as your application's specific wiring, but I understand from the comments that as it stands, a user can cancel a request at the front-end, but no back-end action is taken. (Ie. any back-end response is simply ignored/discarded.)
"Is there a way to make Apache drop cancelled requests?" I'm assuming that your front-end sends the requests directly and without delay to Apache; and onward to PHP > MySQL > PHP > Apache. In that case, no, you can't really have Apache cancel the request that it's already received; or you could hit "stop", but chances are PHP and MySQL are already munching it away...
Holding a "Cancel Window"
However, you could program a "cancel window" lag into your front-end, where requests are only passed on to Apache after e.g. a 0.5-second sleep waiting for a possible cancel. This may or may not have a negative impact on the UX; may be worth implementing to save server resources if a significant portion of requests are canceled. This assumes an UI with Javascript. If you're getting direct HTTP calls to API, you could have a "sleepy proxy receiver" instead.
Using a "Cancel Controller"
How would one cancel PHP/MySQL processes? This is obviously only feasible/doable if calls to your API result in a processing time of any significant duration. If the back-end takes 0.28 sec to process, and user cancels after 0.3 seconds, then there isn't much left to cancel, is there.
However, if you do have scripts that may run for longer, say into a couple of seconds. You could always find relevant break-points in your code, where you have a "not-canceled" check or a kill/rollback routine. Basically, you'd have the following flow:
Front-end sends request with unique ID to main script
PHP script begins the long march for building a response
On cancel: Front-end re-sends the ID to a light-weight cancel controller
Cancel controller logs ID to temporary file/database/wherever
PHP checks at break-points if there's a cancel request for current process
On cancel, PHP executes a kill/rollback routine instead of further processing
This sort of "cancel watch" will obviously create some overhead, and as such you may want to only incorporate this into heavier scripts, to ensure you actually save some processing time in the big picture. Further, you'd only want at most a couple of breakpoints at significant junctions. For read requests, you could just kill the process; but for write requests, you'd probably want to have a graceful rollback to ensure data integrity in your system.
You can also cancel/kill a long-running MySQL thread, already initiated by PHP, with mysqli::kill. For this to make sense, you'd want to run it as MYSQLI_ASYNC, so PHP's around to pull the plug. PDO doesn't seem to have a native equivalent for either async queries or kill. Came across $pdo->query('KILL CONNECTION_ID()'); and PHP Asynchronous MySQL Query (see answer for PDO). Haven't tested these myself. Also see: Kill MySQL query on user abort
PHP Connection Handling
As an alternative to a controller that passes the cancel signal "from the side", you could look into PHP Connection Handling and poll for aborted connection status at your cancel check-points with connection_aborted(). (See "MySQL kill" link above for a code example.)
A CONNECTION_ABORTED state follows if a user clicks the "stop" button in their browser. PHP has a ignore_user_abort() setting, default "Off", which should abort a script on user-abort. (In my experience though, if I have a rogue script and session lock is on, I can't do anything until it times out, even when I hit "stop" in the browser. Go figure.)
If you have "ignore user abort" on false, ie. the PHP script terminates on user abort, be aware that this will be a wholly uncontrolled termination, unless you have register_shutdown_function() implemented. Even so, you'd have to flag check-points in your code for your shutdown function to be able to "rewind the clock" from the termination point onward. Also note this caveat:
PHP will not detect that the user has aborted the connection until an attempt is made to send information to the client. Simply using an echo statement does not guarantee that information is sent, see flush(). ~ PHP Manual on ignore_user_abort
I have no experience with implementing "user abort" over AJAX/JS. For a starting point, see: Abort AJAX Request and Cancel an HTTP fetch() request. Not sure how/if they register with PHP. If you decide to travel down this road, please return and update us with your code / research!
Related
On our PHP website we have noticed, that if I request pages in 2 separate tabs, the second tab will never start loading until the first tab finishes processing the PHP side (uses LiteSpeed webserver). It's a cloud webhosting that should handle a notable number of visitors. We observed the same behavior on previous hosting with Apache.
Could the issue be in webserver configuration? Or could it be something in our application blocking the processing?
If your application uses PHP sessions (intentionally or not), this will generally be the behavior.
When an application calls session_start() it generates a session for you as a user, and will do a file lock on this session until the execution of the site is done, then the lock will be removed, and the next script waiting for the same session can then proceed.
If you're not going to write any data into the session, it's beneficial to call session_write_close() early on, since this will free the lock, and let other processes (Your other tab) allow being processed.
The locking happens because you obviously don't want two processes trying to write to the same file at once, possibly overriding one of the processes data that was written to the session.
Obviously it could also be something else in your application causing this, but you'd have to actually debug the code to figure out. You could possibly use strace on the command-line as well to try to gather more information.
PHP session locking is explained in depth in https://ma.ttias.be/php-session-locking-prevent-sessions-blocking-in-requests/ - it simply goes into a lot more detail than I did above.
If I have a PHP page that is doing a task that takes a long time, and I try to load another page from the same site at the same time, that page won't load until the first page has timed out. For instance if my timeout was set to 60 seconds, then I wouldn't be able to load any other page until 60 seconds after the page that was taking a long time to load/timeout. As far as I know this is expected behaviour.
What I am trying to figure out is whether an erroneous/long loading PHP script that creates the above situation would also affect other people on the same network. I personally thought it was a browser issues (i.e. if I loaded http://somesite.com/myscript.php in chrome and it start working it's magic in the background, I couldn't then load http://somesite.com/myscript2.php until that had timed out, but I could load that page in Firefox). However, I've heard contradictory statements, saying that the timeout would happen to everyone on the same network (IP address?).
My script works on some data imported from sage and takes quite a long time to run - sometiems it can timeout before it finishes (i.e. if the sage import crashes over the weeked), so I run it again and it picks up where it left off. I am worried that other staff in the office will not be able to access the site while this is running.
The problem you have here is actually related to the fact that (I'm guessing) you are using sessions. This may be a bit of a stretch, but it would account for exactly what you describe.
This is not in fact "expected behaviour" unless your web server is set up to run a single process with a single thread, which I highly doubt. This would create a situation where the web server is only able to handle a single request at any one time, and this would affect everybody on the network. This is exactly why your web server probably won't be set up like this - in fact I suspect you will find it is impossible to configure your server like this, as it would make the server somewhat useless. And before some smart alec chimes in with "what about Node.js?" - that is a special case, as I am sure you are already well aware.
When a PHP script has a session open, it has an exclusive lock on the file in which the session data is stored. This means that any subsequent request will block at the call to session_start() while PHP tries to acquire that exclusive lock on the session data file - which it can't, because your previous request still has one. As soon as your previous request finishes, it releases it's lock on the file and the next request is able to complete. Since sessions are per-machine (in fact per-browsing session, as the name suggests, which is why it works in a different browser) this will not affect other users of your network, but leaving your site set up so that this is an issue even just for you is bad practice and easily avoidable.
The solution to this is to call session_write_close() as soon as you have finished with the session data in a given script. This causes the script to close the session file and release it's lock. You should try and either finish with the session data before you start the long running process, or not call session_start() until after it has completed.
In theory you can call session_write_close() and then call session_start() again later in the script, but I have found that PHP sometimes exhibits buggy behaviour in this respect (I think this is cookie related, but don't quote me on that). Obviously, pay attention to the fact the setting cookies modifies the headers, so you have to call session_start() before you output any data or enable output buffering.
For example, consider this script:
<?php
session_start();
if (!isset($_SESSION['someval'])) {
$_SESSION['someval'] = 1;
} else {
$_SESSION['someval']++;
}
echo "someval is {$_SESSION['someval']}";
sleep(10);
With the above script, you will have to wait 10 seconds before you are able to make a second request. However, if you add a call to session_write_close() after the echo line, you will be able to make another request before the previous request has completed.
Hmm... I did not check but I think that each request to the webserver is handled in a thread of its own. Thereby a different request should not be blocked. Just try :-) Use a different browser and access your page while the big script is running!
Err.. I just see that this worked for you :-) And it should for others, too.
Suppose a page takes a long time to generate, some large report for example, and the user closes the browser, or maybe they press refresh, does the PHP engine stop generating the page from the original request?
And if not, what can one do to cope with users refreshing a page a lot that causes an expensive report to be generated.
I have tried this and it seems that it does not stop any running query on the database. But that could be an engine problem, not PHP.
Extra info:
IIS7
MS SQL Server via ODBC
When you send a request to the server, it is executed on the server without any communication with the browser until information is sent back to the browser. When PHP tries to send data back to the browser, it will fail and therefore the script will exit.
However, if you have a lot of code executing before any headers are sent, this will continue to execute until the headers are sent and a failed response is received.
PHP knows when a connection has been closed when it tries to output some data (and fails). echo, print, flush, etc. Aside from this, no, it doesn't; everything else is happening on the server end.
There is little in the way of passing back information about the browser state once a request has been made (or in your case, in progress)
To know if a user is still connected to your site, you will need to implement a long poll / comet or perhaps a web socket.
Alternatively - you may want to run the long query initiated via an ajax call - while keeping the main browser respsonsive (not white screened). This allows you to detect if the browser is closed during the long query with a Javascript event onbeforeunload() to notify your backend that the user has left. (I'm not sure how you would interupt a query in progress from another HTTP request though)
PHP have two functions to control this. set_time_limit(num) able to increase the limit before a page execution "dies". If you don't expand that limit, a page running "too long" will die. Bad for a long process. Also you need ignore_user_abort(TRUE) so the server don't close the PHP process if the server detect that the page has ben closed in the client side.
You may also need to check for memory leaks if you are writing something that use much memory and run for several hours.
When you send a request to the server the server will go away and perform the appropriate actions. IIS/SQL Server does not know if the browser has been closed (and it is not IIS/SQL Server's responsibility to understand this) so it will execute the commands (as told to do so by the PHP engine until it has finished or until the engine kills any transactions). Since your report could be dynamic, IIS will not cache page requests, SQL Server however can cache the last previously ran queries therefore you will see some performance gain from the database backend.
On executing two very simple ajax POST requests (successivly), the Apache server seems to always respond in the same order as they were requested, although the second request takes significant less amount of time to process than the first request.
The time it takes the server to process Request1 is 30 seconds.
The time it takes the server to process Request2 is 10 seconds.
var deferred1 = dojo.xhrPost(xhrArgs1);
var deferred2 = dojo.xhrPost(xhrArgs2);
I expect Apache to achieve some "parallelization" on my dual core machine, which is obviously not happening.
When I execute each request at the same time in a separate broswer then works ok, the Request2 is returned first.
Facts:
httpd.conf has: ThreadsPerChild 50, MaxRequestsPerChild 50
PHP version : 5.2.5
Apache's access log states that both client requests are received at about the same time, which is as expected.
The Php code on the server side is something as simple as sleep(30)/sleep(10)
Any idea about why I don't get the "parallelization" when run from the same browser?
Thanks
When your two requests are sent from the same browser, they both share the same session.
When sessions are stored in files (that's the default), there is a locking mecanism that's used, to ensure that two scripts will not use the same session at the same time -- allowing that could result in the session data of the first script being overwriten by the second one.
That's why your second script doesn't start before the first one is finished : it's waiting for the lock (created by the first script) on the session data to be released.
For more informations, take a look at the manual page of session_write_close() -- which might be a solution to your problem : close the session before the sleep() (quoting) :
Session data is usually stored after
your script terminated without the
need to call session_write_close(),
but as session data is locked to
prevent concurrent writes only one
script may operate on a session at any
time. When using framesets
together with sessions you will
experience the frames loading one by
one due to this locking. You can
reduce the time needed to load all the
frames by ending the session as soon
as all changes to session variables
are done.
Browsers typically have a limit of two connections to the same site (although you may increase that limit in some browsers). Some browsers will keep one connection for downloading things like images etc. and another connection for XHR. Which means that your two XHR calls actually goes out in the same connection, one after the other.
Your browser will return immediately after each XHR call because they are async, but internally it may just batch up the requests.
When you run on two different browsers, obviously they each have the two connections, so the two XHR requests go out in different connections. No problem here.
Now it depends on the browser. If the browser allows you to occupy both connections with XHR calls, then you can get up to two requests running simultaneously. Then it will be up to the server which one to do first.
In any rate, if you try with three (or any number >2) XHR requests simultaneously, you will not get more than 2 executed on the server at the same time on modern browsers.
I want to run a relatively time consuming script based on some form input, but I'd rather not resort to cron, so I'm wondering if a php page requested through ajax will continue to execute until completion or if it will halt if the user leaves the page.
It doesn't actually output to the browser until a json_encode at the end of the file, so would everything before that still execute?
It depends.
From http://us3.php.net/manual/en/features.connection-handling.php:
When a PHP script is running normally
the NORMAL state, is active. If the
remote client disconnects the ABORTED
state flag is turned on. A remote
client disconnect is usually caused by
the user hitting his STOP button.
You can decide whether or not you want
a client disconnect to cause your
script to be aborted. Sometimes it is
handy to always have your scripts run
to completion even if there is no
remote browser receiving the output.
The default behaviour is however for
your script to be aborted when the
remote client disconnects. This
behaviour can be set via the
ignore_user_abort php.ini directive as
well as through the corresponding
php_value ignore_user_abort Apache
httpd.conf directive or with the
ignore_user_abort() function.
That would seem to say the answer to your question is "Yes, the script will terminate if the user leaves the page".
However realize that depending on the backend SAPI being used (eg, mod_php), php cannot detect that the client has aborted the connection until an attempt is made to send information to the client. If your long running script does not issue a flush() the script may keep on running even though the user has closed the connection.
Complicating things is even if you do issue periodic calls to flush(), having output buffering on will cause those calls to trap and won't send them down to the client until the script completes anyway!
Further complicating things is if you have installed Apache handlers that buffer the response (for example mod_gzip) then once again php will not detect that the connection is closed and the script will keep on trucking.
Phew.
It depends on your settings - usually it will stop but you can use ignore_user_abort() to make it carry on.
Depending on the configuration of the web server and/or PHP, the PHP process may, or may not, kill the thread when the user terminates the HTTP connection. If an AJAX request is pending when the user walks away from the page, it is dependent on the browser killing the request (not guaranteed) ontop of your server config (not guaranteed). Not the answer you want to hear!
I would recommend creating a work queue in a flat file or database that a constantly-running PHP daemon can poll for jobs. It doesn't suffer from cron delay but keeps CPU/memory usage to a usable level. Once the job is complete, place the results in the flat file/database for AJAX fetch. Or promise to e-mail the user once the job is finished (my preferred method).
Hope that helps
If the client/user/downloader/viewer aborts or disconnects, the script will keep running until something tries do flush new data do the client. Unless you have used
ignore_user_abort(), the script will die there.
In the same order, PHP is unable to determine if client is still there without trying to flush any data to the httpd.
found the actual solution for my case of it not terminating the connection. The SESSION on my Apache/Php server needed to close before the next one could start.
Browser waits for ajax call to complete after abort.