Upgrade to 1.7.4 and WAPT agent

[corriou]

Good morning,

I manage a WAPT server currently in version 1.7.4 (resulting from successive updates from 1.3=>1.6=>1.7) under Linux for about a hundred workstations under Windows 7 and Windows 10 (Kaspersky Endpoint Protection 11 antivirus on all machines with an identical policy)
Following a server update, many machines are listed as unreachable (or even completely unreachable) and numerous problems have occurred:

1) Creating the agent immediately generated the waptupgrade 1.7.4.6074 package, which is supposed to install. However, although it's marked as installed (and the machines as up-to-date), the console's "WAPT Version" column displays 1.7.3.5 (the previous version of both the server and the agent). Manually installing the agent resolves the display issue.

2) Some powered-on machines are marked as unreachable even though they respond to PINGs from the server. When you connect to one of these machines, the commands "wapt-get"work (wapt-get register -S, wapt-get update...) but the console display remains unchanged. The unreachable machines all have one thing in common: they have opened.. websockets (sometimes dozens) to the server, all in TIME_WAITManually installing the agent does not solve this problem.

3) On the client side, when I run "runwaptservice.bat ldebug", the machine becomes reachable again in the console (and some appear) websockets in ESTABLISHED Or CLOSE_WAITThe only error that appears in the logs is:
C:\Program Files (x86)\wapt\lib\site-packages\urllib3\connection.py:362: SubjectAltNameWarning: Certificate for wapt.XXXXXX has no `subjectAltName`, falling back to

Code: Select all

check for a `commonName` for now. This feature is being removed by major browsers and deprecated by RFC 2818. (See https://github.com/shazow/urllib3/issues/497 for details.)
  SubjectAltNameWarning

As soon as I restart the service in the usual way, she becomes unreachable again.

The machines and the server are set to the same clock via NTP.

If you have any leads, I'd be grateful. Thank you for your help.

[corriou]

I also have, on the server side, lines like:

Code: Select all

2019-05-16 16:01:18,582 WARNING Invalid session 54b83be8349b48729e39afa809e71f03

Edit:
This problem is solved: the agent versions on several machines and the server version differed.

[corriou]

Okay, I'm still having the same problem with the console displaying incorrect information in the "reachable" column. I'm getting all sorts of things:
- machine absent but marked OK
- machine present but marked DISCONNECTED
- machine present but with no reachable status (neither OK nor DISCONNECTED).

Running the `runwaptservice` on the client temporarily fixes the problem. I then see that the WebSocket establishes correctly (the `netstat | findstr wapt` gives me at least one ESTABLISHED entry). As soon as I use the service again, I get a series of WebSocket entries in TIME_WAIT.

I admit I'm out of ideas.

[sfonteneau]

Ah, we've already encountered this issue.

The library that WAPT uses for WebSockets sometimes tends to use a proxy when it shouldn't.

Could you please provide the value of the following file on your computer:


HKEY_USERS\.DEFAULT\Software\Microsoft\Windows\CurrentVersion\Internet Settings\ProxyEnable

[corriou]

Hello,

That's a good point. I just discovered through this key that the proxy address was misconfigured on some machines (an old server). My problem is (almost) solved. I just need to correct it on all the machines.

Thank you very much for your help.

Best regards,