10:41:23 <mj-xmr> selsta, Et al., since today evening I will be available most of the time. If there's something wrong with the CI or some IT stuff needs fixing, please ping me. 10:54:47 <mj-xmr> My plan for the week is to find bottlenecks, less the bottlenecks caused by data generation. 12:00:52 <hyc> macos seems to have a bug re: closing dead connections. ran out of file descriptors because it's not cleaning up. 12:01:20 <hyc> I mean, I'm assuming this is mac-specific because I've never seen a problem like this on linux 12:05:14 <hyc> all of this just further cements my disdain for apple products 12:18:09 <wfaressuissia[m]> "ran out of file descriptors because it's not cleaning up." Have you done any other experiments or it's summary from previous tests ? 12:24:25 <hyc> this is just seeing error messages on console now, with monerod running for a couple days 12:24:40 <hyc> netstat also showed hundreds of conns in CLOSE_WAIT 12:25:42 <hyc> also since macos defaults to nfiles=256, the problem really doesn't take very long to surface 12:27:45 <hyc> was 4 days uptime. had to exit and restart. 12:28:28 <hyc> now I'm seeing a bunch of dead RPC connections accumulate in TIME_WAIT 12:28:43 <hyc> but I think TIME_WAIT is normal... 13:05:58 <wfaressuissia[m]> "netstat also showed hundreds of conns in CLOSE_WAIT", `netstat -tunp` can you paste few such lines ? 14:06:23 <rbrunner> Fun fact: It seems that right now the sum of coinbase transactions on testnet is almost exactly 10 times blockheight: print_coinbase_tx_sum 0 1718900 gives me: 14:06:25 <rbrunner> Sum of coinbase transactions between block heights [0, 1718900) is 17187983.385106828366 consisting of 17185461.956122192283 in emissions, and 2521.428984636083 in fees 16:33:34 <fluffypony> hyc: it uses the same TCP/IP stack as FreeBSD, never had an issue 16:34:02 <selsta> also never had this issue but I don't have incoming connections on Mac