Dead2 commented on issue #1775: Deadlock in shm statistics URL: https://github.com/apache/incubator-pagespeed-mod/issues/1775#issuecomment-394993817 The other proxy is an Nginx one for load-balancing and caching without touching the application server, and we don't want mod_pagespeed at that level. Thus it would have to be a middle-proxy if anything. LoadFromFile is in use now, and no other user than the owner is allowed access to the files/folders as a security measure, thus the need for the whole apache process to change user. But it seems to me that we don't have any concrete evidence that mpm-itk kills a process prematurely. I had a look into the code, but I don't know the apache code well enough to be sure of anything, but from what I saw, there are very few cases where mpm-itk will kill the process (error handling mostly). But if mod_pagespeed can deadlock due to a killed/crashed apache child process, that in itself is not good, is it? Sounds like it needs some way to detect the deadlock and either recover or somehow notify the rest of the system of the situation instead of resulting in a DOS-situation for the whole server. At the very least, a timeout on the lock, and fail the connection instead of waiting endlessly for something that should take <1ms. Ideally it would know what pid holds the lock, and print a warning to that effect to ease debugging, or check whether that pid still exists and remove the lock if the pid is gone.
---------------------------------------------------------------- This is an automated message from the Apache Git Service. To respond to the message, please log on GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org With regards, Apache Git Services
