On 7/18/09, Matthew Toseland <toad at amphibian.dyndns.org> wrote: > On Saturday 18 July 2009 00:00:02 Romain Dalmaso wrote: >> Let's bump this thread, I just break the 1w uptime barrier ! >> >> * Used Java memory: 165 MiB >> * Allocated Java memory: 185 MiB >> * Maximum Java memory: 986 MiB >> * Running threads: 202/5000 >> * Available CPUs: 2 >> * Java Version: 1.6.0_07 >> * JVM Vendor: The FreeBSD Foundation >> * JVM Version: 10.0-b23 >> * OS Name: FreeBSD >> * OS Version: 7.2-RELEASE-p1 >> * OS Architecture: i386 >> >> Store Cache >> Keys 360,043 3,920,557 >> Capacity 3,989,551 3,989,552 >> Data Size 10.9 GiB 119 GiB >> Utilization 9.0% 98.3% >> Read-Requests 2,965,449 2,820,060 >> Successful Reads 145,389 227,792 >> Success Rate 4.9028% 8.0776% > > Ooooh, this is interesting! Your store is getting more than half the success > rate of your cache despite being 10x smaller. Normally stores give very low > success rates; your behaviour is more like what it *should* be ... > >> Writes 13,291 446,364 >> Access Rate 4.88 /sec 4.64 /sec >> Write Rate 0.02 /sec 0.73 /sec >> >> * bwlimitDelayTime: 0ms >> * nodeAveragePingTime: 590ms >> * darknetSizeEstimateSession: 68 nodes >> * darknetSizeEstimate24h: 21 nodes >> * darknetSizeEstimate48h: 38 nodes >> * opennetSizeEstimateSession: 9471 nodes >> * opennetSizeEstimate24h: 5022 nodes > > Woah! Are you a seednode?
Yes, I am. > >> * opennetSizeEstimate48h: 6308 nodes >> * nodeUptime: 1w48m >> * routingMissDistance: 0.0440 >> * backedOffPercent: 4.2% >> * pInstantReject: 0.0% >> * unclaimedFIFOSize: 114 >> * RAMBucketPoolSize: 9.99 MiB / 10.0 MiB >> * uptimeAverage: 100.0% >> >> # Input Rate: 19.0 KiB/s (of 380 KiB/s) >> # Output Rate: 64.8 KiB/s (of 95.0 KiB/s) >> # Total Input: 22.1 GiB (38.2 KiB/s average) >> # Total Output: 35.0 GiB (60.4 KiB/s average) >> # Payload Output: 25.1 GiB (43.3 KiB/sec)(71%) > > Good payload. Output is reasonable but a little disappointing. > > Request success rates? Here they are : Group P(success) Count All requests 4.847% 7,236,216 CHKs 13.308% 2,681,466 SSKs 0.537% 4,554,750 Local requests 4.926% 1,078,807 Remote requests 5.025% 6,157,409 Block transfers 96.543% 541,267 Turtled downstream 55.734% 13,909 Transfers timed out 0.129% 13,909 Turtle requests 52.019% 619 >> >> On 6/21/09, xor <xor at gmx.li> wrote: >> > >> > I have "nodeUptime: 1w1d" right now, so let's talk about statistics, >> > they >> > have >> > changed recently: >> > >> > # opennetSizeEstimateSession: 10233 nodes >> > # opennetSizeEstimate24h: 4736 nodes >> > # opennetSizeEstimate48h: 6299 nodes >> > # routingMissDistance: 0.0792 >> > >> > # Total Input: 21.5 GiB (31.8 KiB/s average) >> > # Total Output: 48.1 GiB (71.2 KiB/s average) >> > >> > Detailed timings (local CHK fetches) >> > Successful 15.295s >> > Unsuccessful 12.661s >> > Average 12.766s >> > >> > Success rates >> > Group P(success) Count >> > All requests 5.151% 6,773,032 >> > CHKs 15.057% 2,058,489 >> > SSKs 0.513% 4,714,543 >> > Local requests 4.897% 1,121,112 >> > Remote requests 5.068% 5,651,920 >> > Block transfers 96.351% 496,477 >> > Turtled downstream 87.984% 10,794 >> > Transfers timed out 0.574% 10,794 >> > Turtle requests 58.887% 2,048 >> > >> > If you're interested in more stats just ask. >
