• John Walker posted an update in the group Group logo of UpdatesUpdates 3 weeks, 5 days ago

    2019 July 24

    The weird errors continued overnight, coming also from IPv6
    address REDACTED.  Since whois is
    useless on IPv6, I can't tell if this is the same user.

    Restarted: systemctl restart php-fpm to see it that will make a difference. After the restart, it reports available memory 2,919,156 kB.

    Nope--just got another one: [Wed Jul 24 12:00:36.266919 2019] [proxy_fcgi:error] [pid 2887:tid 140051715032832] [client REDACTED:52679] AH01071: Got error 'PHP message: PHP Fatal error: Allowed memory size of 134217728 bytes exhausted (tried to allocate 20480 bytes) in /server/pub/ on line 4155\n', referer:

    OK, what appears to be happening is that php-fpm is hitting its memory limit, which is configured at the default of 128 Mb. This setting is in /etc/php-fpm.d/www.conf: ;php_admin_value[memory_limit] = 128M I changed it to: php_admin_value[memory_limit] = 192M to see what will happen. I restarted php-fpm after changing the configuration.

    Well, yes, and here we go again.... [Wed Jul 24 12:18:45.647071 2019] [proxy_fcgi:error] [pid 2887:tid 140051715032832] [client REDACTED:52810] AH01071: Got error 'PHP message: PHP Fatal error: Allowed memory size of 201326592 bytes exhausted (tried to allocate 7853692 bytes) in /server/pub/ on line 178\n', referer: . However, the rate at which the errors are occurring appears to have been reduced.

    All right, let's try going to: php_admin_value[memory_limit] = 256M in /etc/php-fpm.d/www.conf. Several down-market references: . . . claim this is the minimum for reliable operation in WordPress.

    It's been ten hours since I increased the memory_limit to 256 Mb and we haven't had a single out of memory failure, so it looks like this bloodsucker has been staked. Why it started happening yesterday, after 18 months of the site's running without a single incident, 2 months since the last WordPress update, and only for two users, remains a deep mystery.