Kirill Korotaev <dev at sw.ru> writes: > We have not seen any degradation here in real cases, > but probably you are right and pid hash can be allocated taking into account > physical memory as it is done for TCP/ip/other hashes? It is but it is currently capped at 4K entries. With 4K entries and 32K pids our worst case is usage is a hash chain 9 entries long. At 4M pids our hash chains are 1000 entries long, which sucks. > But not sure, it is worth bothering right now... Maybe it worth first to make > some > simple test, say: > > 1. run 50,000 tasks. > 2. run some benchmark > > and compare benchmark results with different hash sizes? > What do you think? If it is easy sure. The real point of where things degrade is past 50K processes though. The practical question is if systems using containers are using noticeably more pids than anyone else. So far the responses I have gotten indicate that users aren't. So at least until we descend into multi-core madness it sounds like the current structures are fine, but it might be worth moving the cap on the number of pid hash table entries at some point in the future. Eric