Author

Topic: [4+ EH] Slush Pool (slushpool.com); Overt AsicBoost; World First Mining Pool - page 1066. (Read 4382653 times)

donator
Activity: 2772
Merit: 1019
I'm having trouble connecting to mining.bitcoin.cz:80, chrome tells me ""

This has been the case for about 10 hours now, my ip is/was: 85.176.112.67

Weirdly enough, I can connect through tor (using exit 46.19.138.242)

Any explanation?

Can you try a TCP traceroute?  I don't know if such a tool is available for Windows, but there is one for Linux.  It will essentially do a traceroute with TCP packets by setting the SYN packet's TTL to 1 initially and incrementing it for every packet sent thereafter, in an attempt to discover where the traffic is really going (using the ICMP TTL exceeded errors to determine the path, similar to how regular traceroute works).

This is an effective technique for discovering transparent proxies or other bizarre/evil things done by your ISP or any router between you and the destination.

Thanks for the hint. Unfortunately I can access the web-page again now :|
member
Activity: 98
Merit: 10
SYN flooding should be easy for the provider to handle.

...which is also already solved on the server, so this is not an issue.

Your server is still getting hit with the SYN packets ... why not have your ISP stop them and give your server a break [this is the sort of thing that is best blocked by a router, not a production server].  I assume you used IPTABLES on Linux to block syn attacks, but that still leaves your network bandwidth utilized and your CPU (Kernel) dealing with each of these.  Your choice, but handling attacks on the server when a router would do a better job [in this case] may prove to be resource expensive as you grow.
legendary
Activity: 1386
Merit: 1097
SYN flooding should be easy for the provider to handle.

...which is also already solved on the server, so this is not an issue.
member
Activity: 98
Merit: 10
My point is that they don't like a DOS attack passing through their network either and will often stop it at their routers.  The advantage is that their network is not strained, they have removed the traffic going into your network and taking up your resources and they also [usually] have a better idea how to best stop the attack and not just filter it.

How do they know which traffic is needed and which is not? Currently SYN flooding and similar is solved by pool server itself and the rest of possible attacks are usually application related, so I don't know how Linode can help with this...

SYN flooding should be easy for the provider(s) to handle.  Application specific attacks you are correct about; you must make sure they are not vulnerable and then if necessary filter [at the ISP if possible].  If it is just some script kiddie, then a complaint against the source IP pool owner will often put a halt to it.  You did mention it was coming from a single block if I remember correctly.  This isn't my area of expertise, so I don't have much to offer beyond that.
legendary
Activity: 1386
Merit: 1097
My point is that they don't like a DOS attack passing through their network either and will often stop it at their routers.  The advantage is that their network is not strained, they have removed the traffic going into your network and taking up your resources and they also [usually] have a better idea how to best stop the attack and not just filter it.

How do they know which traffic is needed and which is not? Currently SYN flooding and similar is solved by pool server itself and the rest of possible attacks are usually application related, so I don't know how Linode can help with this...
member
Activity: 98
Merit: 10
You have business class hosting through your ISP, right?  Can't they get a handle on this for you?

Handle what? IP blacklist? I'm completely OK with managing this blacklist for myself.

To connection problems - I don't think it is problem on the pool side, because it works without problems for two thousands of workers, only two or three people reported some troubles. It definitely looks like problem of their firewalls, routers etc. I once had similar problems with my router box - I had high packet loss, slow roundtrips. Restarting helped. It's common problem of cheap router boxes...

My point is that they don't like a DOS attack passing through their network either and will often stop it at their routers.  The advantage is that their network is not strained, they have removed the traffic going into your network and taking up your resources and they also [usually] have a better idea how to best stop the attack and not just filter it.

Just my thoughts. 
legendary
Activity: 1386
Merit: 1097
You have business class hosting through your ISP, right?  Can't they get a handle on this for you?

Handle what? IP blacklist? I'm completely OK with managing this blacklist for myself.

To connection problems - I don't think it is problem on the pool side, because it works without problems for two thousands of workers, only two or three people reported some troubles. It definitely looks like problem of their firewalls, routers etc. I once had similar problems with my router box - I had high packet loss, slow roundtrips. Restarting helped. It's common problem of cheap router boxes...
member
Activity: 98
Merit: 10
Plese send me your IPs to PM. It's possible that they fall into IP blacklist because of server flooding - I'll check it.

You have business class hosting through your ISP, right?  Can't they get a handle on this for you?

legendary
Activity: 1386
Merit: 1097
Getting a 502 Bad Gateway error on the main page right now, even though my miners can connect fine. Hmmm.

Was fixed in few minutes. I really have to recompile bitcoind for webserver tomorrow...
full member
Activity: 178
Merit: 100
Certified fox posing as a cat posing as a human
Getting a 502 Bad Gateway error on the main page right now, even though my miners can connect fine. Hmmm.
legendary
Activity: 3920
Merit: 2349
Eadem mutata resurgo
Traceroute can do a "TCP traceroute", just add -P TCP -p 80.

For more info use "man traceroute".

actually he needs to do it for port 8332, i.e. miner port
linux command

$sudo traceroute mining.bitcoin.cz -P TCP -p 8332
member
Activity: 116
Merit: 10
Traceroute can do a "TCP traceroute", just add -P TCP -p 80.

For more info use "man traceroute".
hero member
Activity: 499
Merit: 500
Traceroute and tracert are ICMP trace route utilities not TCP trace route.

I suggest searching on "TCP traceroute" and find what is available for your OS.
member
Activity: 116
Merit: 10
Under Linux it is called traceroute and under Windows tracert.

Istead of telling people about SYN-Flags, TTL and TCP you could just link to wikipedia or somewhere.
full member
Activity: 182
Merit: 107
I'm having trouble connecting to mining.bitcoin.cz:80, chrome tells me ""

This has been the case for about 10 hours now, my ip is/was: 85.176.112.67

Weirdly enough, I can connect through tor (using exit 46.19.138.242)

Any explanation?

Can you try a TCP traceroute?  I don't know if such a tool is available for Windows, but there is one for Linux.  It will essentially do a traceroute with TCP packets by setting the SYN packet's TTL to 1 initially and incrementing it for every packet sent thereafter, in an attempt to discover where the traffic is really going (using the ICMP TTL exceeded errors to determine the path, similar to how regular traceroute works).

This is an effective technique for discovering transparent proxies or other bizarre/evil things done by your ISP or any router between you and the destination.
legendary
Activity: 1386
Merit: 1097
Any explanation?

No, this IP isn't in blacklist and I don't know about any connection troubles on server side...
donator
Activity: 2772
Merit: 1019
I'm having trouble connecting to mining.bitcoin.cz:80, chrome tells me ""

This has been the case for about 10 hours now, my ip is/was: 85.176.112.67

Weirdly enough, I can connect through tor (using exit 46.19.138.242)

Any explanation?
legendary
Activity: 1386
Merit: 1097
Maybe slush recently made a change to not allow connections to worker account if one is already "connected"?

No, more miners using same worker account should not lead to connection troubles...
donator
Activity: 2772
Merit: 1019
I'm having trouble connecting with more then 1 worker. I have four instances using the same worker on one computer - they are all working fine. A worker on a different computer won't connect. If I turn off the other worker it will.

This just started last night.

Last I checked you had to create a worker account for each worker.

Maybe slush recently made a change to not allow connections to worker account if one is already "connected"?
legendary
Activity: 1386
Merit: 1097
Plese send me your IPs to PM. It's possible that they fall into IP blacklist because of server flooding - I'll check it.
Jump to: