Hi There, I'm currently having an issue and was hoping for a little advice on workarounds or ways to solve it.
We have the need to extract a lot of information out of memcached before serving requests, I thought I had something solved by using a WSGIDaemonProcess that my wsgi scripts can use. And then simply load all the memcached data into a singleton object at the module level of the wsgi scripts. And this all seem to work perfectly, it takes around 40 seconds to load all of the data but works fine after that. My problem comes when attempting to update the wsgi files, or trying apachectrl graceful command. At this point all subsequent requests are having to wait for the 40 load of data to complete. Our current deployment is using mod_wsgi 3.4, but I've updated to 4.5 and have the same issue. Is there a way for mod_wsgi to only start tearing down processes after there are new processes ready and fully loaded to handle requests? The only option I can see to move forward is to load the data required in a background thread. This way requests can still be handled, but they may have to load a section of the data required from memcached at request time. Not the ideal situation as this is what I'm trying to move away from with the pre-loading of this data. But at least this way it will only happen if a process goes down, or someone manually gracefully restarts apache. Would really appreciate some advice on how best to move forward, I did try to search in this forum for a similar question, but couldn't quite find the correct search terms. Thank you Steve -- You received this message because you are subscribed to the Google Groups "modwsgi" group. To unsubscribe from this group and stop receiving emails from it, send an email to modwsgi+unsubscr...@googlegroups.com. To view this discussion on the web visit https://groups.google.com/d/msgid/modwsgi/bf2e4b18-458f-43e3-9130-363f1d524335o%40googlegroups.com.
