It seems i've found a bug in SFS 1.6.18 patch.
This patch has got new feature:
- Added more severe policy for malformed packets (XML, JSON and string based). Malformed requests will cause immediate client disconnection.
The bug itself:
In case such malformed packet arrives, SFS drops it and closes the socket but does not release a user "slot" assigned for it.
It leads to max. available user room overflow (for e.q for a trial SFS version - only 20 simultaneously users available) and SFS stops accept new connections.
SFS Pro 1.6.18 patch bug
Re: SFS Pro 1.6.18 patch bug
Thanks, I've opened a ticket with your report.
We're investigating and I'll report back soon.
Stay tuned.
We're investigating and I'll report back soon.
Stay tuned.
Re: SFS Pro 1.6.18 patch bug
UPDATE: we're not sure how to reproduce this.
Can you please send us a proof of concept of how to recreate the issue? You can use our support@... email box.
Thanks
Can you please send us a proof of concept of how to recreate the issue? You can use our support@... email box.
Thanks
Re: SFS Pro 1.6.18 patch bug
I would have preferred you sent us an email.
If there's a way to attack the server it's best not to describe how it's done step by step in a public forum
Anyways thanks for the details, we'll investigate and let you know.
If there's a way to attack the server it's best not to describe how it's done step by step in a public forum
Anyways thanks for the details, we'll investigate and let you know.
-
- Posts: 27
- Joined: 20 Jun 2016, 18:30
Re: SFS Pro 1.6.18 patch bug
It also happens to my server
But this happens only if the client sends a lot of requests to the extension so the server kicking the room visits
But this happens only if the client sends a lot of requests to the extension so the server kicking the room visits
Last edited by ransaymour on 27 Jan 2017, 17:12, edited 2 times in total.
Re: SFS Pro 1.6.18 patch bug
We're working on a patch to be release next week.
I'll post more details here as we progress.
cheers
I'll post more details here as we progress.
cheers
-
- Posts: 27
- Joined: 20 Jun 2016, 18:30
Re: SFS Pro 1.6.18 patch bug
Thank you
Re: SFS Pro 1.6.18 patch bug
UPDATE:
Here is patch 1.6.19
Please download it and apply to your local environment, then see if you can still reproduce the problem.
From our end the problem seems solved.
Thanks
Here is patch 1.6.19
Please download it and apply to your local environment, then see if you can still reproduce the problem.
From our end the problem seems solved.
Thanks
Re: SFS Pro 1.6.18 patch bug
Thanks for the patch, I've applied it, but I can't confirm that the problem is completely solved.
At the scenario that I described early, (where the are two sockets are involved) the SFS now closes the connected socket from the malformed message come and not performs login procedure following by the malformed message, but NOT IN ALL CASES. When this happens very fast in a loop-cycle, the valid user authentication procedure still take in place eventually (may be for 1 in 100-200 iterations, and the available users room storage is still could be overflowed .
At the scenario that I described early, (where the are two sockets are involved) the SFS now closes the connected socket from the malformed message come and not performs login procedure following by the malformed message, but NOT IN ALL CASES. When this happens very fast in a loop-cycle, the valid user authentication procedure still take in place eventually (may be for 1 in 100-200 iterations, and the available users room storage is still could be overflowed .
Re: SFS Pro 1.6.18 patch bug
I am not sure how to reproduce this, but I have another consideration and it is about the lack of a login validation.
Without a login credential check you're leaving the door open for abuse, because the potential attacker can spam the server knowing that any user will be accepted.
To me this seems like the bigger issue in the scenario you have described.
In any case if you have a script or proof of concept that can reproduce the issue we'll be happy to look into this further.
Without a login credential check you're leaving the door open for abuse, because the potential attacker can spam the server knowing that any user will be accepted.
To me this seems like the bigger issue in the scenario you have described.
In any case if you have a script or proof of concept that can reproduce the issue we'll be happy to look into this further.
-
- Posts: 1
- Joined: 26 Feb 2018, 12:38
Re: SFS Pro 1.6.18 patch bug
SmartFoxServer Pro is constantly crashing. I've the error logs where am I supposed to send it @Lapo?
Re: SFS Pro 1.6.18 patch bug
Hi,
you can start a new thread and provide all the details about the problem you're seeing.
Here's a list of what we need (it's for SFS2X but applies to PRO as well):
viewtopic.php?f=18&t=16497
Thanks
you can start a new thread and provide all the details about the problem you're seeing.
Here's a list of what we need (it's for SFS2X but applies to PRO as well):
viewtopic.php?f=18&t=16497
Thanks
Re: SFS Pro 1.6.18 patch bug
prapat50 wrote:SmartFoxServer Pro is constantly crashing
Please open a new thread and give us the details.
This will help you:
viewtopic.php?f=18&t=16497
Return to “SmartFoxServer 1.x Discussions and Help”
Who is online
Users browsing this forum: No registered users and 85 guests