38200+ entries in 0.334s

a111: Logged on 2016-12-30 16:28 mircea
_popescu: but he might be interested to hear about it.
mircea_popescu: you can also set bgwriter
_lru
_maxpages to 0 and disable background writing altogether
mircea_popescu: eg bgwriter
_delay you may want to be set low for this reason.
mircea_popescu: and work
_mem should prolly be larger than 4mb, but hard to guess how large without a profile, and this is a major resource consumption ticket, so you actually want to do some maffs.
mircea_popescu: 1) shared
_buffers is to be per spec "25% of available ram" ; but it does diminish returns in the gb. you probably have it as 128mb, make it 2gb say.
mircea_popescu: incidentally asciilifeform since we're now doing open source db optimization shared
_buffers is probably a larger concern. what is it ? defaults to 125mb but i'd readily see it 1-4gb in your case.
a111: Logged on 2016-12-30 14:32 Framedragger: (do note, 'work
_mem' is per user / per request. so may be easier to DoS. thought i should mention this for completeness)
a111: Logged on 2016-12-30 06:58 ben
_vulpes: felipelalli, vexare, wyrdmantis, Sinclair6, luke-jr please fix your bouncers
mircea_popescu: (they did decide to move over to unity last year, then they abandoned the plan.
_ mircea_popescu views with mind's eye diana
_coman 's beard growing inches/second in the minds of alf
Framedragger: mircea
_popescu: because you could tell postgres to flush rows (forcing all caching layers) every 100 rows, not every 1 row as currently specified
Framedragger: mircea
_popescu: consider a scenario in which you knew how much data you could lose ("up to 100 last rows"), and you could check if you lost any (last row id == last-id-processed.txt ? false : true). that being said, this way things become more wibbly-wobbly, so probably fuck that. :(
Framedragger: "Note that open
_sync writing is buggy on some platforms (such as Linux), and you should (as always) do plenty of tests under a heavy write load to make sure that you haven't made your system less stable with this change. Reliable Writes contains more information on this topic. " oh god. more inserts/sec but zero data loss => probably can't help you much. documentation doesn't encourage me :/
Framedragger: here's what i'm thinking: disable synchronous
_commit , but set 'checkpoints' so that results are flushed to db every $n inserts/updates. i can see however how you may barf from such an idea, "it's either reliable, or isn't".
Framedragger: (do note, 'work
_mem' is per user / per request. so may be easier to DoS. thought i should mention this for completeness)
☟︎ Framedragger: ah hm. tbh i'd still change work
_mem because it's ridoinculously low by default, but i hear ya.
Framedragger: work
_mem (used for in-memory sorts) is 4 MB default. 4 MB. (9.5 anyway). set it to 50 MB as per advisory at least.
Framedragger: mircea
_popescu: it's dark here in the northern hemisphere, god it's deperessing :( mornin'..
Framedragger: right. i just thought about checkpoint
_completion
_target (set to say 0.9) which may help with inserts, but ultimately you're right, physical reality
ben_vulpes: i believe that mod6 has a solid one as well, pete
_dushenski's has been blackholed of late
davout: ben
_vulpes: any suggestions for such a node?
phf: ben
_vulpes: you mean like ~parse~ the output of patch?
ben_vulpes: ^^ mircea
_popescu asciilifeform mod6 trinque and any other vtronicists pls to opine
phf: ben
_vulpes: where'd that 0 come from?
phf: ben
_vulpes: well, we're kind of constrained by the hardcode -p1 behavior, but i've no idea if that's an implementation detail or a spec
trinque: mircea
_popescu: yep works now
trinque: mircea
_popescu: ah you know what, excluding nicks with no ratings breaks this, because writing the keyfile happens inside that logic.
trinque: ben
_vulpes: and yes, app code needn't be anywhere near www
trinque: there's a listening process consuming these pg
_notify events, of the form (gen-nick-page "trinque") which debounces them according to some sane interval, i.e. if a few updates hit the same user in a short span, it will result in one single rebuild
trinque: ben
_vulpes and I had an interesting conversation yesterday about how to handle static sites generated from a db. idea we ended up with was that we'd have triggers which emit a pg
_notify signal when the "dirty bit" has flipped for any page.
trinque: mircea
_popescu: it'll be up there in a sec, let ya know
a111: Logged on 2016-12-29 23:32 ben
_vulpes:
http://btcbase.org/log/2016-12-28#1591573 << the diff line is distinct from the --- / +++ lines, does one ever see a patch file where the files compared aren't prefixed with a/ or b/ ?
a111: Logged on 2016-12-28 02:39 phf:
http://btcbase.org/patches/hashes_and_errors#L118 you don't really want to do this. you're subseq'ing there to strip the a/ b/ but that's not at all a guarantee! i have a vpatch with `diff -ib -ruN /Users/pf/cmucl20d-build/src/hemlock/abbrev.lisp src/abbrev.lisp` in it for example. at the very least you want to abstract it away into its own function. that would correctly operate on a hashed-path datastructure.
pete_dushenski: mircea
_popescu: have done some, will try more banning.
mircea_popescu: pete
_dushenski looks like stock blackholing. ipban the offenders, see what happens.
jurov: ben
_vulpes: prb 0.11 and newer insist on using improved protocol
a111: Logged on 2016-12-29 22:50 ben
_vulpes: run moar macos or how does it go
ben_vulpes: pete
_dushenski: you have to reboot the whole machine?
pete_dushenski: mircea
_popescu: i don't disagree from a philosophical standpoint but nor can i tolerate having dead fucking trb nodes. that i should have to reboot a machine ~daily~ is the death of bitcoin. yukoners never had it so bad.
mircea_popescu: pete
_dushenski there's really no good reason to not serve arbitrart txn data. the fact that "modern" prb nodes can't support this is entirely their doom.
a111: Logged on 2016-12-09 02:38 ben
_vulpes: asciilifeform should probably poop more often
davout: ben
_vulpes: it's even worse here, it's to *exit* a position
davout hands ben
_vulpes some wine
ben_vulpes: pete
_dushenski: asking or being asked?
a111: Logged on 2015-07-02 20:44 ascii
_field: 'getdata is used in response to inv... ...t can be used to retrieve transactions, but only if they are in the memory pool or relay set - arbitrary access to transactions in the chain is not allowed to avoid having clients start to depend on nodes having full transaction indexes (which modern nodes do not).'
a111: Logged on 2016-12-29 21:04 ben
_vulpes: mod6: moreover the general case is looking to be "project-genesis.vpatch"
pete_dushenski: mircea
_popescu: i have sincerely nfi why my node is asking for these. it's at ~full height.
mircea_popescu: pete
_dushenski why is your node asking for these ? or is it ?
shinohai: mircea
_popescu will be saddened to know the inventor of the red Solo cup has died.