Followup question: How would someone measure the current use of thread_pool_workspace?
On Sat, Nov 23, 2013 at 8:59 PM, Jason Price <[email protected]> wrote: > I'm working with vmod_redis to do some metrics gathering and sundry. > But I think I've got a memory leak in there. > > Is there a recent, canonical implementation of vmod_redis? I've > worked my way through github a couple of times trying to find 'the > one', but to no avail. (ideally, the vmod_redis implementation would > support the pipeline call) > > My symptoms: Varnish, vmod_redis, and redis all work beautifully, but > after several hours, the redis connections start returning "Out of > memory" errors. The server has stacks (10+GB) of unused ram, so > that's not the problem. I believe my threads are bumping into > 'thread_pool_workspace" or a similar limit, but I'm having challenges > proving that assertion. I know redis isn't the problem because other > clients can connect to it without issue at the same time varnish > can't. > > Thank you for any pointers; > Jason _______________________________________________ varnish-misc mailing list [email protected] https://www.varnish-cache.org/lists/mailman/listinfo/varnish-misc
