Attack on DevShares / client misbehaving?

[wackou]

looks like the misbehaving client has left the network at 00:00 UTC last night (times on the plot are UTC+1):

[wackou]

ok, thanks. Will report again when the new devshares is out, then. OTOH bitshares 0.7.0 is running very smoothly and the memory leak is gone, great job!

[vikram]

It's very strange and we don't really know what's going on. We are working to get DVS 0.8 out as soon as we can which has lots of changes; if the behavior still persists with the latest code then we will have to investigate more closely.

[wackou]

an update on the situation:

last night the cpu usage on the spike has been slowly fading out:



however, today it starting going up a lot more again, and nearly maxes out 1 core now:



(it doesn't show anything before ~15:20 because I was running it on another machine at that time)

Maybe it is related to the current fork with the init delegates somehow?

[wackou]

my devshares delegate (first screenshot of this thread, this post: https://bitsharestalk.org/index.php?topic=14852.msg192315#msg192315) is running 0.6.3 on debian jessie, 2-core xeon and 4G RAM.

the second and third screenshot are from my laptop which is a 2012 macbook pro, with osx 10.9, devshares 0.6.3 as well.

here's a new one from my laptop where I completely removed the devshares dir:



as you can see, it starts by synchronizing the blockchain and then the same pattern appears. Note that total average cpu usage is not really that high, it's more the pattern that bothers me (so for instance just checking your average daily usage of your server will not show any unusual cpu usage). I am pretty sure also that this appeared 3 days ago and wasn't present before.

Unrelated to this issue, but still worth noticing, is that my BitShares delegate has an extremely low cpu usage (uncannily low, actually), at the cost of a small memory leak (this appeared in bts 0.6.2, 0.6.1 had a more normal cpu usage and no mem leak). Here's a screenshot (specs are the same as my dvs delegate, 2-core xeon and 4G RAM)

[vikram]

What OS's are you running under, and what's the specs (CPU, memory) for the machines you're running on?

And also what version DevShares are you running?

[liondani]

This is like in "contact".. alien message.. :D

lol  +5%

[dannotestein]

Any opinion from core devs on the matter? This has been going on for more than 48 hours now...

I looked at my devshares delegate on Linux and also at my devshares GUI client, and they seem to be operating within normal parameters (both are less loaded as to cpu/memory usage than my respective bitshares clients). My devshares delegate is consuming around 800MB, my bitshares delegate is consuming around 1GB.

My devshares delegate is being sent invalid blocks from the same peer as you are. This is most likely just an out-of-date client from before a hard-fork.

What OS's are you running under, and what's the specs (CPU, memory) for the machines you're running on?

[Ander]

This is like in "contact".. alien message.. :D

Alien message from Come from Beyond, trying to test our system. ;)

[jsidhu]

probably CfB

[xeroc]

paging @vikram !

[wackou]

Any opinion from core devs on the matter? This has been going on for more than 48 hours now...

[38PTSWarrior]

This is like in "contact".. alien message.. :D

[wackou]

yes, apart from that, everything is running smoothly, as usual. But given the looks of it, I doubt that it's a real attack, more like a bug that is somehow causing this.

[xeroc]

@come-from-beyond: is this you?

is the client running as expected?