archive snowblossom slack general
prev 2018-10-23 next
2018-10-23 01:56:41
alistar
@Clueless Arktika 1.4.0 said INFO: Error talking to mining pool: io.grpc.StatusRuntimeException: UNAVAILABLE: Channel shutdownNow invoked, but 1.2.0 is okay to mining
2018-10-23 02:01:37
Fireduck
1.4.0 arktika works fine for me
2018-10-23 02:02:32
alistar
So I dont know why it is happened. I must use 1.2.0 again.
2018-10-23 02:04:00
Fireduck
Sure nothing else is different?
2018-10-23 02:04:12
Fireduck
None of that network code has changed
2018-10-23 02:05:09
alistar
nothing different, same configure files? is configure files different between 1.20 and 1.4.0?
2018-10-23 02:07:57
Fireduck
Same config file?
2018-10-23 02:09:27
alistar
find error.
2018-10-23 02:09:45
alistar
I need update my Pool application.lol
2018-10-23 02:22:51
Clueless
@alistar what pool are you connected to?
2018-10-23 02:23:05
alistar
my own pool
2018-10-23 03:54:35
Fireduck
well, glad it is nothing I need to fix. :wink:
2018-10-23 15:27:01
Fireduck
@Joko I still like you but was just cleaning up perms
2018-10-23 15:30:24
Joko
Haha I like you too mate :smile:
2018-10-23 15:32:04
Fireduck
yeah, you had github commit access
2018-10-23 16:12:40
Joko
Damn, I should have abused my powers when I had my chance!
2018-10-23 16:20:19
Clueless
So, nicely, you can use `cpulimit` to throttle your miner if you want, or perhaps an iolimit
2018-10-23 17:06:32
Clueless
@Rotonen I'm able to get 750+K mining in a VM, really not bad.
2018-10-23 17:07:01
Clueless
using `cpulimit` to throttle to `15%` for testnet nets me `Oct 23 10:06:30 vmsnow1 cpulimit[27821]: INFO: Mining rate: 19456.636/sec - at this rate 0.487 hours per block`
2018-10-23 17:45:53
Fireduck
Anyone want to make a testnet faucet so I don't have to?
2018-10-23 17:46:12
Fireduck
My plan is to encourage people to use testnet to play with things (like the upcoming Android client)
2018-10-23 17:57:49
mjay
@Fireduck I´ll take the job
2018-10-23 17:58:56
Fireduck
@mjay awesome
2018-10-23 19:49:15
mjay
First step: Testnet explorer online. http://testnet.snowblossom-explorer.org/
2018-10-23 19:49:54
Fireduck
awesome. Let me know when you want some testnet coin
2018-10-23 19:51:11
mjay
You can send me some: snowtest:varv2n46rmmf79n5etad2lw8flx758axx0gd4pum
2018-10-23 19:54:05
Fireduck
done
2018-10-23 19:56:12
mjay
thank you
2018-10-23 20:14:24
Rotonen
@Clueless rather use chrt if you want it to stay out of the way
2018-10-23 21:15:46
mjay
Is there a minimal length for snow addresses?
2018-10-23 21:16:10
Fireduck
not really
2018-10-23 21:16:29
Fireduck
if they end up with a bunch of zeros on the front the base32 form will be smaller
2018-10-23 21:16:45
Fireduck
What language are you working with? We can probably come up with a validator.
2018-10-23 21:17:09
mjay
I´ll just do some basic validation, should be fine
2018-10-23 21:17:20
Fireduck
cool
2018-10-23 21:17:27
Fireduck
the send will reject it if it is nonsense
2018-10-23 21:17:30
Fireduck
so whatever
2018-10-23 21:18:05
mjay
contains only base32, is 40 chars or shorter, it should be fine
2018-10-23 21:18:20
mjay
snowtest: removed before test
2018-10-23 21:18:50
Fireduck
the address can be valid with snowtest: on the front or not
2018-10-23 21:19:07
Fireduck
when presenting anything to a user, you should always have snowtest: on
2018-10-23 21:19:09
Fireduck
for clarity
2018-10-23 21:19:37
mjay
I remove it before the check if its present, and always send it without to the node
2018-10-23 21:20:14
Fireduck
cool. I'm just saying you can send to the node either way.
2018-10-23 21:43:07
mjay
5 testsnow each time
2018-10-23 21:49:40
Fireduck
1) awesome
2018-10-23 21:49:44
Fireduck
2) I've filled it up more
2018-10-23 21:50:03
mjay
thanks
2018-10-23 21:50:22
mjay
once it drops below 1k snow a miner will start
2018-10-23 21:51:17
Fireduck
that is fancy
2018-10-23 21:54:16
Rotonen
cronjobs are
2018-10-23 21:56:13
Fireduck
hey, that is like an if statement and something to not spawn a new one each time the job runs (presumably)
2018-10-23 21:56:23
Fireduck
and maybe even some mechanism to turn it off when not needed
2018-10-23 21:56:25
Fireduck
so fancy
2018-10-23 21:58:12
mjay
its a simple cronjob. Every 15 minutes: kill the miner (nothing happens if none is running), and start a new one if balance < 1000
2018-10-23 21:59:49
mjay
still untested, so I expect it to be broken
2018-10-23 21:59:58
Rotonen
not even doing a sigterm timeout dance?
2018-10-23 22:00:34
Fireduck
I like it, it has a certain brutal efficency
2018-10-23 22:01:08
Fireduck
walk into room. Shoot one foot above guard desk. If a guard is required, call temp agency for a new one.
2018-10-23 22:01:45
mjay
efficient shooting using SIGTERM
2018-10-23 22:02:11
mjay
SIGKILL sounded a little harsh
2018-10-23 22:40:20
Rotonen
@mjay no i mean handling the fact signal sending is blocking
2018-10-23 22:46:07
mjay
you mean, waiting until the process stops?
2018-10-23 22:54:41
Rotonen
in general handling an indefinite stall upon SIGTERM
2018-10-23 22:55:32
mjay
I just added another SIGKILL 5 seconds after the SIGTERM
2018-10-23 22:56:47
Rotonen
how does the script get there if the kill is blocking
2018-10-23 22:57:13
mjay
It never happened to me that a kill blocks
2018-10-23 22:58:23
Rotonen
seems you mostly deal with well behaving well written software
2018-10-23 22:58:53
mjay
how can you get it to block?
2018-10-23 23:00:05
mjay
I had processes that catch sigterm and ignore it, zombies, cuda-processes with gpu crashed (ignoring even sigkill)
2018-10-23 23:00:19
mjay
but kill never blocked (so far)
2018-10-23 23:00:36
Rotonen
i'm sorta paranoid on that sorta unixy stuff, so i use fancy bells and whistles the kernel gives me to decide my course of action, like https://lwn.net/Articles/288056/ Like most versions of Unix, Linux has two fundamental ways in which a process can be put to sleep. A process which is placed in the TASK_INTERRUPTIBLE state will sleep until either (1) something explicitly wakes it up, or (2) a non-masked signal is received. The TASK_UNINTERRUPTIBLE state, instead, ignores signals; processes in that state will require an explicit wakeup before they can run again.
2018-10-23 23:02:07
Rotonen
i think you can also get a file descriptor on the signal you've sent these days, but that's too new for me
2018-10-23 23:04:59
Rotonen
@mjay if you want to implement a test program for yourself, see this https://cr.yp.to/docs/selfpipe.html seek out the classic book, and misimplement the example
2018-10-23 23:06:16
mjay
From Mr. Bernstein, nice
2018-10-23 23:24:51
mjay
Those old unix books are fun to read and experiment with
2018-10-23 23:28:01
Rotonen
yeah, funny how that parallels with dusty magic tomes for wizardry
2018-10-23 23:30:21
Fireduck
I like that djb posts on the post quantum mailing list I'm on
2018-10-23 23:30:41
Fireduck
I just watch, no real clue what those guys are on about
2018-10-23 23:32:17
mjay
what mailing list is this?
2018-10-23 23:33:34
Fireduck
Yep
2018-10-23 23:35:00
Fireduck
Watching with an eye towards adding winners to snowblossom
2018-10-23 23:35:29
mjay
Nice, joined
2018-10-23 23:35:53
Rotonen
but yeah, people with a lifetime of experience in this stuff debating how to deal with unknown unknowns is probably... dense
2018-10-23 23:41:28
Fireduck
hehe
2018-10-23 23:41:29
Fireduck
it is
2018-10-23 23:49:01
Fireduck
There was some brilliant shade last week
2018-10-23 23:49:18
Fireduck
Someone announced an attack on one of the candidates
2018-10-23 23:49:43
Fireduck
DJB came back a few days later saying the attack had no merit.
2018-10-23 23:49:56
Fireduck
:fire:
2018-10-23 23:54:51
mjay
Its really impressive to read through these papers, barely understanding it, and to know someone came up with this stuff
2018-10-23 23:55:42
Fireduck
Specialization is great
2018-10-23 23:56:07
Fireduck
Those folks are specialist in that area
2018-10-23 23:58:27
Fireduck
I specialize in doing dumb things with hashed tries