Repository navigation
Accounts gets wasted by dead proxies #2402
Description
Today in our < 24h ban phase we need every account we have and we need to save them as much as possible.
If a proxy gets offline or is unreachable for a scan then the proxy will be removed from the live list. But also the account is going into -ari mode. If you have a .csv file with thousand of accounts and a proxy dies you will waste a lot of fresh accounts.
[ WARNING] Account XXXXXX proxy 113.IP.IPI.IPIP:3128 is not in a live list any more. Switching accounts...
Expected Behavior
If the proxy is not alive any more:
- Kick it out of the proxy list
- Give the affected account a new proxy and go on scanning
Current Behavior
A dead proxy grabs a new accounts from your pool and waste them all after a long time.
Possible Solution
Change code :)
Steps to Reproduce (for bugs)
- Start scanning
- Wait until a proxy dies
- Grep the log
Activity
crhbetz commentedon Dec 21, 2017
-asi and -ari are not account-management tools. They were meant to simulate human-like behaviour as far as playing time goes to try to increase account lifetime, but it's now established knowledge that it's most efficient to just have your accounts run 24/7 until they don't work anymore.
urinella commentedon Dec 21, 2017
@crhbetz Yes, I think you misunderstood.
Example:
Map start and scanning maybe 12 hours:
Problem:
Now you wasted 10 accounts because the map is using 260 accounts in the first 12 hours instead of 250.
crhbetz commentedon Dec 21, 2017
I get that this is not the expected behaviour when a proxy goes offline. But as things are right now, it's also just not a good idea to use an account file that carries more workers than you need at any given time.
In your example, just use 4 files of 250 accounts each, set rest interval to 0, and cycle your file once per day. This can be easily accomplished with a simple script in the language of your choice.
friscoMad commentedon Dec 21, 2017
Why not reducing -ari to the minimun and reducing proxy check interval can also help.
Anyway scanner accounts should work for 3 days unless you stress them a lot.
urinella commentedon Dec 22, 2017
@friscoMad You didn't read my comment. Please read it first before you spread out your 'knowledge'.
Then you know that your comment doesn't make sense...
And 3 days is no way. Even just login with new accounts getting them banned after 24 hours.
I own 12k accounts and tried everything in every constellation.
friscoMad commentedon Dec 22, 2017
@urinella
Just by login fresh validated accounts, without ip or domain flagged does not cause banning in 24h I have been using the same account for 5 days for minor api tests (just login the full 38 requests and maybe a gmo and a spin several times a day) and only yesterday it did receive the warning maybe it is SB from day 3 but for sure not temp banned.
Accounts in my maps have been working for 3 full days before SB (not tempbanning) since SB came back a couple of months ago and that's the same reports I get from other mappers not that I just got lucky.
About my "knowledge" it is true I thought that the account recicler did put the accounts back on top of the queue but it is not the case so it does not help for you use case.
The correct solution would be to change the proxy and keep the account working but it is not easy to do with the current code.
sebastienvercammen commentedon Dec 30, 2017
@urinella You can tone down the attitude. @friscoMad was helping. He's also one of our core devs, so if you're looking for help these are not the people to show a bad attitude.
That being said, the request is valid and we should consider the response from the request and disable a proxy instead of just requesting new accounts when we can determine it was the proxy's fault.