"Undefined ENS memory problem" ???

[groch]

hi.
2day i was twice bombarded by such message
http://gfx.neostrada.pl/ens.gif
it happen while there was heavy port redirection activity
and lots of frozen icmp connections
still server didnt crash or anything and internet activity was keept alive
should i care about this??
Groch

[MattP]

Hi,

Which version are you running? We fixed this version in 5.2 so if you haven't already, please upgrade. This is related to a memory leak in the ENS driver so you should try to correct it.

Regards,

[groch]

its 5.2.2 version

[groch]

and one another thing, when redirecting ports there are lots of other computers (IP's) listed in the activity tab it makes preaty big mess,
is there a way that redirected connectons would be child objects of the computer they are redirected to, casue now there are soo many other IP listed i have trouble finding my own computers on the list :)
regards.
Groch

[webmistress32]

I just upgraded to 5.2.2 and am experiencing errors in the system log.

will the 5.2.3 upgrade solve this?

thanks. Laura

[Pascal]

What errors are you seeing in the system log ? There are four messages logged in succession which are reported as Errors but are actually just information messages. (Traffic statistics, memory in use, etc.)

[webmistress32]

the error is in the system log and it is repeated over and over and over.

the error is "UNDEFINED ENS MEMORY PROBLEM" and then something about "possible memory exhaustion". it's the same error message in each entry in the system log.

I don't have the machine available while at work. I can get more info when I'm at home, if you like.

Laura

ps sorry somehow I created a new thread, too :-( this is a new forum format for me.

[webmistress32]

here is the exact error message, about every second in the log.

Undefined ENS memory problem - possible cause memory exhaustion. Statistics: Kerner mem: 0, Frag mem: 0

[genie]

Aye, looks like there might be a reason for this particular message - did people seeing this message run Wingate under very heavy load? The reason of it can be the shortage of network buffers available for the next packet to be sent. This problem can be solved by changing the number of preallocated network buffers for Wingate driver (registry value HKLM/System/CurrentControlSet/Services/QbikHk<xxx>/Parameters/MaxNetBuffers) to 1024 or higher (up to 8192) - reboot is required for this change to take an effect.

[webmistress32]

I have surfed over to the hklm ... services ... qbikxxx ... parameters and there is no maxnetbuffers.

I am happy to add the value but what data type shall I make it?

thanks. Laura

[genie]

Sorry for not mentioning it- yes, add it (type DWORD).

[luisnina]

gretings
this value DWORD, 1024 -8169 must be hexadecimal or decimal?

thanks.

[adrien]

those numbers are decimal ones.

Regards

Adrien