client crashed at end

[ctm]

Dealing 15000 30000 Chowaha

deadhead 🤮 blinds 8000
NUT-Z blinds 15000
1     NUT-Z            400 15000  [  ] [  ]
2 B > deadhead 🤮  36600  8000  [  ] [  ]
                    My cards are: [Ah] [9s]
It is 7000 to you
deadhead 🤮 raises 15000 to 30000
NUT-Z calls and is all-in

Nothing else appeared afterward. JavaScript console had a ton of errors:

 No route matched, provide a redirect prop to the router to handle cases where no route can be matched
En @ 1.mb2-web.js:1
13:37:54.524 wasm://wasm/011a9756:1 Uncaught RuntimeError: unreachable
    at wasm-function[9301]:0x44ed24
    at wasm-function[8397]:0x44b905
    at wasm-function[8376]:0x44b6a1
    at wasm-function[3084]:0x3ad6c2
    at wasm-function[333]:0x19473c
    at wasm-function[4177]:0x3ebd67
    at wasm-function[1770]:0x31950c
    at wasm-function[2734]:0x390846
    at wasm-function[7122]:0x43bd11
    at wasm-function[7140]:0x43c29f
13:37:54.528 wasm://wasm/011a9756:1 Uncaught RuntimeError: unreachable
    at wasm-function[9301]:0x44ed24
    at wasm-function[8280]:0x44a9b7
    at wasm-function[7140]:0x43c281
    at wasm-function[7735]:0x444a84
    at wasm-function[7085]:0x43b15f
    at wasm-function[7037]:0x43a22d
    at wasm-function[3994]:0x3e3a10
    at wasm-function[7169]:0x43cb2e
    at wasm-function[8809]:0x44d51b
    at wasm-function[7309]:0x43f0ba
13:37:54.528 wasm://wasm/011a9756:1 Uncaught RuntimeError: unreachable
    at wasm-function[9301]:0x44ed24
    at wasm-function[8280]:0x44a9b7
    at wasm-function[7140]:0x43c281
    at wasm-function[7735]:0x444a84
    at wasm-function[7085]:0x43b15f
    at wasm-function[7037]:0x43a22d
    at wasm-function[3994]:0x3e3a10
    at wasm-function[7169]:0x43cb2e
    at wasm-function[8809]:0x44d51b
    at wasm-function[7309]:0x43f0ba
13:37:54.528 wasm://wasm/011a9756:1 Uncaught RuntimeError: unreachable

FWIW, I wasn't able to reach the network when I went to report this, so there's a chance that a weird network disconnect caused this, but Nut-Z saw weirdness on his side of things too, but he didn't get a screen capture and I'm not yet logging enough info to look into it. I did go onto devctm.com and used docker logs but didn't see anything notable:

[root@devctm ~]# docker logs mb2
docker logs mb2
2020-03-21 02:38:09,093 INFO  [mb2] Started 0.0.0.0:8080
2020-03-21 02:38:09,097 INFO  [actix_server::builder] Starting 1 workers
2020-03-21 02:38:09,097 INFO  [actix_server::builder] Starting "actix-web-service-0.0.0.0:8080" service on 0.0.0.0:8080
2020-03-21 02:38:09,097 INFO  [actix_server::builder] Starting "actix-web-service-0.0.0.0:8081" service on 0.0.0.0:8081
request: ClosedConnection(TableSessionId(1968658759757913917))
request: ClosedConnection(TableSessionId(8057065356568942593))
request: ClosedConnection(TableSessionId(3865829141073594042))

My recollection is that the ClosedConnection message only occurs when there is an improper shutdown of a connection but without timestamps I can't even be sure that one of them was mine.

[ctm]

I had a similar panic today at the end, although this time I was watching memory and it happened right as the Chrome Task Manager showed we ticked over from 1.9 gig to 2.0 gig. I think we started out at 168 meg.

So, we definitely have a memory leak, but I'm going to put all the details in #112 for now (even though that's about CPU rather than memory).

[ctm]

I believe this is different from the memory leak, because I've seen it at the end several times. It does not happen every time, more like once every four or so times. Most recently it happened while I was spectating and memory had already gone over 2.1GB, so it wasn't a matter of it hitting the "magic" 2GB.

I suspect a dev build would tell us exactly what is going wrong, but simply not stripping symbols might be sufficient. I should have done that before.

[ctm]

2 GB is actually largely a magic number. I learned that while spending a lot of time tracking down #112. As such, I'm closing this, but I can reopen if needed.