Out of Buffers (error msg)

[mkerridge]

Help! I am running wingate 3.0.5 (i know its old) we are getting ready to upgrade, however, until then I need this working properly.
1. When I open the enging to log in after I enter pswd I get
"no sockets" "out of buffers" (This is on the server machine)

2. No person can access the internet - on the local machine - it says"Out of Buffers"

Thanks

Maria

[genie]

Yes, this out-of-buffers situation happened quite often in the old versions. The best solution for you would be to upgrade to higher version of Wingate.

[LFrank]

Genie - obviously not only with the old one, unfortunately.

Was just going to post an issue I'm seeing since a week or so... before my PC was rebooted at least once a day due to ongoing development - these are finished now and the PC is "ON" constantly.

Or, should be "ON" - now I started to encounter "Out of Buffers" as reported above after a bit more than 24h. I switched of the "logging" from my Intel Pro network cards and it improved. But only for 36 to 48h. After that, I again get that message, Inetinfo is reaching 100%CPU (but can now be restarted and connections are still possible) but opening Gatekeeper claims the buffers...

Any idea what to try else on 6.01?

Thanks,

Lutz

[adrien]

Hi

What service pack and OS are you running? Having problems like that with InetInfo.exe makes me wonder whether there is an OS bug.

If you look in Task Manager in such cases, and choose to view handles, how many handles does it show WinGate using?

Adrien

[LFrank]

I'm running XP pro on SP2 patch-level - I didn't have the handle display activated, will feed back as soon as it happens again.

You're most certainly right - it happened to occur after finishing the SP2 testphase (which forced a lot of frequent reboots)

Thanks for the feedback,
Read You soon

Lutz

[LFrank]

36h later ...

Wingate shows 849 connections at the time I got to the PC showing the error

[LFrank]

Any news on those 6.01 - Out of Buffer Errors?? Or any idea what I could provide to get it solved (it really happens on regular 24 to 36h basis).

I'm just going to load an Image of the OS drive which was taken around or just before - the issue became really nasty. Perhaps this 'straw' helps...

CU

Lutz

[adrien]

849 connections is a lot of connections.

If you go to the command prompt, and type "netstat -an" do you see a lot of connections in there, or many connections in the TIMED_WAIT state?

It is possible that some connections are not being closed properly.

There may be some TCP/IP registry settings you can tweak to give your system more sockets, but I wonder why there are so many connections. Are these all active?

Adrien

[LFrank]

Had a few tests to run which forced reboots earlier than 24h.

Waiting again for about 36h yesterday its back - strange thing is that I even can't run nbtstat or netstat commands any longer - the will crash with an exception. Only remedy - reboot.

Sigh