session_start(): Unable to clear session lock record

[Murugan-k]

Hi,

We configured session store as memcached, it was working fine in PHP 5.6.X,
after upgraded to PHP 7.0.30, when we start session its throwing error as "session_start(): Unable to clear session lock record", below are the configuration we are using,

PHP Version 7.0.30
memcached 3.0.3
libmemcached version 1.0.18
ZendFramework 2.4.13

below are the memcached settings configured in ini,

memcached.sess_lock_expire => 0 => 0
memcached.sess_lock_max_wait => not set => not set
memcached.sess_lock_retries => 5 => 5
memcached.sess_lock_wait => not set => not set
memcached.sess_lock_wait_max => 2000 => 2000
memcached.sess_lock_wait_min => 1000 => 1000
memcached.sess_locking => 1 => 1
memcached.sess_number_of_replicas => 0 => 0
memcached.sess_persistent => 0 => 0
memcached.sess_prefix => memc.sess. => memc.sess.

also we have tried with below configuration changes,

memcached.sess_lock_retries => 10 => 10 // 50 also tried
memcached.sess_lock_wait_max => 150 => 150
memcached.sess_lock_wait_min => 150 => 150
session.lazy_write => Off => Off

but it didn't work, keep on getting "session_start(): Unable to clear session lock record"

[woolardfa]

Refer to issue #269. Specifically this comment and this comment.

You see those messages because the initial page request process is finishing up after an extended period of time, and the session it is trying to remove has already been removed by an secondary call, such as an ajax call, using the same session state/cookie.

[Murugan-k]

Hi @woolardfa ,

As i said before i have already changed the lock_wait_min and max to 150, also increased the retry count to 2000. Now we haven't received session_start(): Unable to clear session lock record

But now we are getting the empty response, also getting the below error

Error in `/usr/sbin/httpd': free(): invalid pointer: 0x00007ffb105e0400

[sodabrew]

You’ll need to get a backtrace from the httpd process with debugging symbols installed (often a separate distribution package like apache-https-dbg), otherwise there is no way to see where the crash came from.

Alternatively if you can set up a minimum repro case in a small script. Let’s say it’s the number of retires. If 1000 retries never crashes but 3000 retries always crashes, and a small script (without any of your application code) can trigger the crash, then I can run that on my own local machine and debug the code.

Last bit - file a new ticket with backtrace or repro script for the crash please.

[woolardfa]

Made this commented also in #269 -- Just an FYI... have tested PHP 7.2 (7.2.6), the underlying issue where session_start returns true when it in fact failed to acquire the session lock, and subsequently corrupts the existing session data, has been fixed.

A call to session_start() now returns false when it failed to acquire the session lock.

[cappadaan]

+1, i still have this problem using PHP 7.4.0

[vukitoso]

Ubuntu 18.04
php5.6-fpm/bionic,now 5.6.40-15+ubuntu18.04.1+deb.sury.org+1 amd64
php-memcached/bionic,now 3.1.4+2.2.0-1+ubuntu18.04.1+deb.sury.org+20191129 amd64

php.ini
session.save_handler = memcached
session.save_path = "l27.0.0.1:11211"

Warning: session_start(): Write of lock failed in /var/www/sitedd.local/1.php on line 2
Warning: session_start(): Unable to clear session lock record in /var/www/sitedd.local/1.php on line 2

[ideerge]

I disabled lazy_write, works like before

[geltix]

We had the same issue and we found that this is not a real error: This happens if you connect asyncronous to your website that opens a session with session_start() and you open another page while the first page has not been closed before. It seems that like this php does not have the posibble change session values and therefore it shows this errors. We made following to solve: The asynchronous website call via javascript we passed variables via get and in the php page we did not open session_start().