--- Day changed Wed Nov 20 2019 00:27 -!- roshii [~yaaic@93-238-145-85.ftth.glasoperator.nl] has quit [Ping timeout: 240 seconds] 00:44 -!- reallll is now known as belcher 02:14 -!- simon_ [~simon@93-238-145-85.ftth.glasoperator.nl] has joined #joinmarket 02:34 -!- azizLIGHT [~azizLIGHT@unaffiliated/azizlight] has joined #joinmarket 07:07 < simon_> back on my "not filling orders" issue: 07:07 < simon_> after moving bitcoin wallet to ramdisk, i had yi-gen running for a night: it did try to fill 3 orders but again, it was too slow to answer with something 30 to 35s per wallet query 07:07 < simon_> i then understood that my system was just short of memory and filling in the swap file reflecting on ramdisk access time... 07:08 < simon_> i did therefore shut most of other services running on my system to free up memory and had the yi-gen running for another half day: one order was filled by did not went thru once again with 27s delay per query... 07:08 < simon_> after making sure memory wasn't an issue anylonger, i did mount my docker partition without btfrs copy on write feature, which i understand leads to read performance degradation when doing lots of random writes 07:08 < simon_> these setting proved to be successful since i had one order filled with wallet requests being completed in 18s 07:13 < belcher> so it was the swap writing that did it? 07:38 -!- ghost43 [~daer@gateway/tor-sasl/daer] has quit [Ping timeout: 260 seconds] 07:39 -!- technonerd [~techno@gateway/tor-sasl/technonerd] has quit [Ping timeout: 260 seconds] 07:40 -!- kristapsk [~KK@gateway/tor-sasl/kristapsk] has quit [Ping timeout: 260 seconds] 07:43 -!- ghost43 [~daer@gateway/tor-sasl/daer] has joined #joinmarket 07:43 -!- technonerd [~techno@gateway/tor-sasl/technonerd] has joined #joinmarket 07:59 < simon_> most probably yes: with a 2 gb swap file, volume activity did spike when accessing bitcoin wallet even if it was on the ramdisk that implied up to 40 s wait time 07:59 < simon_> this 40s is just the same as if wallet.dat was on a disk 08:01 < simon_> the 18s i experience now is similar to the 20s i had when first testing ramdisk - which at time of test had not moved the data to swap 08:31 -!- simon_ [~simon@93-238-145-85.ftth.glasoperator.nl] has quit [Quit: simon_] 08:40 -!- simon_ [~simon@93-238-145-85.ftth.glasoperator.nl] has joined #joinmarket 09:02 -!- roshii [~yaaic@93-238-145-85.ftth.glasoperator.nl] has joined #joinmarket 09:06 -!- johnonchain [~johnoncha@2001:41d0:fc63:9c00:2931:aec:9468:a5e7] has joined #joinmarket 09:21 -!- johnonchain [~johnoncha@2001:41d0:fc63:9c00:2931:aec:9468:a5e7] has quit [Ping timeout: 250 seconds] 09:23 -!- takamatsu [~takamatsu@unaffiliated/takamatsu] has joined #joinmarket 09:28 < waxwing> from "obtained tx" INFO msg to ">> sig" DEBUG message i seem to pretty much always have < 0.1 seconds. there is a delay like 15-20 seconds between the fill action and the receipt of the tx, but that's normal because the taker is waiting to receive data from all makers, and it is throttled on IRC (i.e. we have to deliberately delay it). 09:28 < waxwing> we have had situations where people with hard disks were getting big delays from bitcoin core wallet queries, but that was only really a problem when we were importing 60 addresses at once AND it was only a problem because of a current bug in Bitcoin Core (database lock) 09:29 < waxwing> but that is not currently the case, there is no 60 address import, it should import two addresses and should not take meaningful time (perhaps a second or two?) 09:29 < waxwing> simon_, ^ 09:35 -!- undeath [~undeath@hashcat/team/undeath] has joined #joinmarket 10:03 -!- kristapsk [~KK@gateway/tor-sasl/kristapsk] has joined #joinmarket 11:18 < belcher> yes 18 seconds is still a huge amount 12:09 < simon_> well, my system is actually reporting 60 address imports, that is exactly what takes 18s 12:09 < waxwing> which commit are you on according to `git log` simon_ ? 12:10 < simon_> i'm using release 0.5.5 12:10 < simon_> let me check what git log says... 12:11 -!- undeath [~undeath@hashcat/team/undeath] has quit [Quit: WeeChat 2.6] 12:12 < waxwing> simon_, right. it seems we forgot that that 60 address import is in that release, but i can see, it is. 12:12 < waxwing> can you do `git pull origin master` (if necessary do `git stash` first to preserve any local changes to your yg script) 12:13 < waxwing> and try using master. it should remove that largish delay (which only happens with slow HDD and again, only happens because there is a bug in Bitcoin Core which will be fixed either in 0.19 or 0.20) 12:15 < simon_> ok thanks, i will adjust my container to git pull from master rather than wget/tar on release 12:16 < simon_> i'll report back when i have some data 12:24 < belcher> if the 60 address commit is the problem then why wasnt the delay solved when he put wallet.dat on ramdisk 13:07 < waxwing> i think he said it swapped because there was insufficient resources (mem) 13:50 < belcher> maybe its a combination of several factors 15:21 -!- roshii [~yaaic@93-238-145-85.ftth.glasoperator.nl] has quit [Ping timeout: 265 seconds] 17:37 -!- AgoraRelay [~jmrelayfn@p5DE4A6AE.dip0.t-ipconnect.de] has quit [Ping timeout: 240 seconds] 17:37 -!- AgoraRelay [~jmrelayfn@p5DE4A76A.dip0.t-ipconnect.de] has joined #joinmarket 17:38 -!- CgRelayBot [~CgRelayBo@p5DE4A6AE.dip0.t-ipconnect.de] has quit [Ping timeout: 265 seconds] 17:39 -!- CgRelayBot [~CgRelayBo@p5DE4A76A.dip0.t-ipconnect.de] has joined #joinmarket 18:50 -!- MaxSan [~four@209.58.183.78] has joined #joinmarket 19:08 -!- slivera [slivera@gateway/vpn/privateinternetaccess/slivera] has joined #joinmarket 19:26 -!- reallll [~belcher@unaffiliated/belcher] has joined #joinmarket 19:29 -!- belcher [~belcher@unaffiliated/belcher] has quit [Ping timeout: 245 seconds] 20:47 -!- MaxSan [~four@209.58.183.78] has quit [Quit: Leaving.] 22:18 -!- adlai [~adlai@unaffiliated/adlai] has quit [Read error: Connection reset by peer] 22:20 -!- adlai [~adlai@unaffiliated/adlai] has joined #joinmarket