{"id":72251,"date":"2020-03-03T02:42:20","date_gmt":"2020-03-02T23:42:20","guid":{"rendered":"https:\/\/prohoster.info\/blog\/vypusk-obrabotchika-nehvatki-pamyati-earlyoom-1-4"},"modified":"2020-03-03T16:10:14","modified_gmt":"2020-03-03T13:10:14","slug":"vypusk-obrabotchika-nehvatki-pamyati-earlyoom-1-4","status":"publish","type":"post","link":"https:\/\/prohoster.info\/en\/blog\/news\/vypusk-obrabotchika-nehvatki-pamyati-earlyoom-1-4","title":{"rendered":"The earlyoom memory shortage handler 1.4 has been released.","gt_translate_keys":[{"key":"rendered","format":"text"}]},"content":{"rendered":"

After eight months of development has been published<\/a><\/noindex> the release of the background process earlyoom 1.4<\/a><\/noindex>, which periodically checks the amount of available memory (MemAvailable, SwapFree) and attempts to proactively respond to memory shortages. The project code is written in C and is distributed<\/a><\/noindex> is under the MIT license.<\/p>\n

If the amount of available memory is below a specified value, earlyoom forcibly (by sending SIGTERM or SIGKILL) terminates the process that is consuming the most memory (the one with the highest value in \/proc\/*\/oom_score), without allowing the system to reach a state where its buffers are cleared and hindering swap operation (the OOM (Out Of Memory) handler in the kernel is triggered when the memory shortage condition has already reached critical levels and usually by that point the system is unresponsive to user actions).<\/p>\n

Earlyoom supports sending notifications about forcibly terminated processes to the desktop (using notify-send), and also provides the ability to define rules where process names can be specified using regular expressions for which termination is preferable (the «—prefer» option) or whose stopping should be avoided (the «—avoid» option). <\/p>\n

Key changes<\/a><\/noindex> in the new release: <\/p>\n