> Dnia wtorek, 29 listopada 2011 11:47:35 Roberto De Ioris pisze: >> > Hi, >> > >> > I'm doing some stress test of uWSGI with 3 servers: >> > >> > nginx -> uwsgi fastrouter -> 2x uwsgi backend >> > (nginx and fastrouter are on same server) >> > >> > and one backend gets 2x more requests. I've tested it with apache >> > benchmark >> > using: >> > >> > ab -n 10000 -c 1 -H "Host: app.domain.com" >> > "http://$nginx_address/$some_url"; >> >> Ok, just committed the fix for round robin and the support for gathering >> statistics from the fastrouter. >> >> Some info for more "obscure" field: >> >> reference: is the amount of currently running requests for this node >> >> death_mark: if 1 this node is dead but there are still requests hanging >> on >> it (it will be completely deleted after the last requests will timeout) >> >> last_check: the last timestamp on which the node has been announced > > It works well and round robin not works, each node got 500 requests. But > during first test I've managed to put fastrouter into infinite loop, > unfortunately I can't reproduce it so I'll just discribe how I got this > bug: > > 1. I've upgraded uWSGI, added fastrouter-stats and restartet both workers > and > fastrouter > 2. I've started ab as before > 3. immediately I've run curl "fasterouter_ip:fastrouter-stats_port" and > I've > got: > > { "version": "1.0-dev-1732", > "uid": 0, > "gid": 0, > "cwd": "/", > "fastrouter": ["=0","=0"], > "subscriptions": [ > { "key": "app.domain.com", > "hits": 13, > "nodes": [ > {"name": "172.16.200.55:3001", "modifier1": 0, > "modifier2": 0, "last_check": 1322564749, "requests": 7, "tx": 36762, > "ref": > 1}, > {"name": "172.16.200.56:3001", "modifier1": 0, > "modifier2": 0, "last_check": 1322564749, "requests": 6, "tx": 36762, > "ref": 0} > ] > }, > { "key": "pingapp.local", > "hits": 2, > "nodes": [ > {"name": "172.16.200.55:3000", "modifier1": 0, > "modifier2": 0, "last_check": 1322564634, "requests": 1, "tx": 34, "ref": > 0}, > {"name": "172.16.200.56:3000", "modifier1": 0, > "modifier2": 0, "last_check": 1322564633, "requests": 1, "tx": 34, "ref": > 0} > ] > }, > { "key": "app.domain.com", > "hits": 13, > "nodes": [ > {"name": "172.16.200.55:3001", "modifier1": 0, > "modifier2": 0, "last_check": 1322564749, "requests": 7, "tx": 36762, > "ref": > 1}, > {"name": "172.16.200.56:3001", "modifier1": 0, > "modifier2": 0, "last_check": 1322564749, "requests": 6, "tx": 36762, > "ref": 0} > ] > }, > { "key": "pingapp.local", > "hits": 2, > "nodes": [ > {"name": "172.16.200.55:3000", "modifier1": 0, > "modifier2": 0, "last_check": 1322564634, "requests": 1, "tx": 34, "ref": > 0}, > {"name": "172.16.200.56:3000", "modifier1": 0, > "modifier2": 0, "last_check": 1322564633, "requests": 1, "tx": 34, "ref": > 0} > ] > }, > { "key": "app.domain.com", > "hits": 13, > "nodes": [ > {"name": "172.16.200.55:3001", "modifier1": 0, > "modifier2": 0, "last_check": 1322564749, "requests": 7, "tx": 36762, > "ref": > 1}, > {"name": "172.16.200.56:3001", "modifier1": 0, > "modifier2": 0, "last_check": 1322564749, "requests": 6, "tx": 36762, > "ref": 0} > ] > }, > [it goes a looooong way down repeating all apps all the time] > > uWSGI fastrouter instance was eating cpu all the time. Maybe there is some > race condition in the fastrouter-stats code and I've hit time window where > is > occurs. > >
It is incredible how a feature added for a scope (monitoring) became so kick-ass in another one (debugging). Thanks, i have enough material to debug the problem. I will let you know soon. -- Roberto De Ioris http://unbit.it _______________________________________________ uWSGI mailing list [email protected] http://lists.unbit.it/cgi-bin/mailman/listinfo/uwsgi
