[c-nsp] con0 and XRemote problem which ends with serious memory issues

[Andy B.]

Hi,

I'm facing an issue with a 6500 running SXI5 that eventually ends up
eating all memory and reload is the only way to solve it:

#who
    Line       User       Host(s)              Idle       Location
   0 con 0                XRemote: 24 clients  01:54:29


This box has been up for 12 hours, and the number of XRemote increases
over the day, while no console is attached to con 0.

I tried to clear line con 0 to no avail and then I also tried to
identify the TCB by using sh tcp brief and then clearing the TCB.

Here is an example:

5B09CF34  x.x.189.1.8000          x.218.199.147.2055          CLOSED

#sh tcp tcb 5B09CF34
Connection state is CLOSED, I/O status: 8, unread input bytes: 9
Mininum incoming TTL 0, Outgoing TTL 255
Local host: x.x.189.1, Local port: 8000
Foreign host: x.218.199.147, Foreign port: 2055

Enqueued packets for retransmit: 0, input: 1  mis-ordered: 0 (0 bytes)

Event Timers (current time is 0x289A8DC):
Timer          Starts    Wakeups            Next
Retrans             1          0             0x0
TimeWait            0          0             0x0
AckHold             1          1             0x0
SendWnd             0          0             0x0
KeepAlive           0          0             0x0
GiveUp              0          0             0x0
PmtuAger            0          0             0x0
DeadWait            1          0       0x291DA40

iss: 2341161453  snduna: 2341161454  sndnxt: 2341161454     sndwnd:  65535
irs: 1925376677  rcvnxt: 1925376687  rcvwnd:       4119  delrcvwnd:      0

SRTT: 52 ms, RTTO: 1968 ms, RTV: 1916 ms, KRTT: 0 ms
minRTT: 416 ms, maxRTT: 416 ms, ACK hold: 200 ms
Flags: passive open, higher precedence, retransmission timeout
  path mtu capable

Datagrams (max data segment is 1460 bytes):
Rcvd: 4 (out of order: 0), with data: 1, total data bytes: 9
Sent: 4 (retransmit: 0), with data: 0, total data bytes: 0


(Note that all those XRemote sessions seem to be on port 8000, but I
cannot explain why)

#clear tcp tcb 5B09CF34
[confirm]
 [OK]


It does not disappear from the list and I tried to clear it numerous
times, and eventually it disappeared.



Furthermore, while this goes on, this is spamming my logs:

May 12 15:32:05.660 CEST: %SYS-2-GETBUF: Bad getbuffer, bytes= -10
-Process= "Exec", ipl= 0, pid= 119
-Traceback= 42377590 422007D0 422013A4 42200A60 42203984 4052E5B0
4233FCAC 4055C178 41392EC8 41392EB4
May 12 15:32:05.928 CEST: %SYS-2-GETBUF: Bad getbuffer, bytes= -10
-Process= "Exec", ipl= 0, pid= 119
-Traceback= 42377590 422007D0 422013A4 42200A60 42203984 4052E5B0
4233FCAC 4055C178 41392EC8 41392EB4
May 12 15:32:06.180 CEST: %SYS-2-GETBUF: Bad getbuffer, bytes= -10
-Process= "Exec", ipl= 0, pid= 119
-Traceback= 42377590 422007D0 422013A4 42200A60 42203984 4052E5B0
4233FCAC 4055C178 41392EC8 41392EB4
May 12 15:32:06.432 CEST: %SYS-2-GETBUF: Bad getbuffer, bytes= -10
-Process= "Exec", ipl= 0, pid= 119
-Traceback= 42377590 422007D0 422013A4 42200A60 42203984 4052E5B0
4233FCAC 4055C178 41392EC8 41392EB4
May 12 15:32:06.684 CEST: %SYS-2-GETBUF: Bad getbuffer, bytes= -10
-Process= "Exec", ipl= 0, pid= 119
-Traceback= 42377590 422007D0 422013A4 42200A60 42203984 4052E5B0
4233FCAC 4055C178 41392EC8 41392EB4
May 12 15:32:06.936 CEST: %SYS-2-GETBUF: Bad getbuffer, bytes= -10
-Process= "Exec", ipl= 0, pid= 119
-Traceback= 42377590 422007D0 422013A4 42200A60 42203984 4052E5B0
4233FCAC 4055C178 41392EC8 41392EB4
May 12 15:32:07.200 CEST: %SYS-2-GETBUF: Bad getbuffer, bytes= -10
-Process= "Exec", ipl= 0, pid= 119
-Traceback= 42377590 422007D0 422013A4 42200A60 42203984 4052E5B0
4233FCAC 4055C178 41392EC8 41392EB4



The main problem is that the number of XRemote sessions is going up to
128 and it is slowly eating up all available memory until it is all
used up and you are forced to reload. Last time this happened, memory
was full in roughly 6 weeks.

I have no service and no ACL using port 8000.

When I reload the box it's good for a while, and then it starts over again.


Has anyone seen this behaviour? How can this be solved without
reloading every once and a while?

Thanks.

Andy