SUMMARY: (sort of) Virtual Adrian, "High Retransmissions"

From: Stephen Harris (sweh@mpn.com)
Date: Mon Apr 06 1998 - 03:27:27 CDT


My original question asked about SE3.0 and queried what virtual_adrian
meant when it said...

Adrian detected TCP/IP problem (amber): Thu Apr 2 15:31:02 1998
Moderate retransmissions
    tcpIn tcpOut InConn/s OutConn/s %Retran ListenDrop/s Reset/s Attempt/s
     1499 2692 0.07 0.30 22.29 0.00 0.03 0.00

Adrian detected TCP/IP problem (red): Thu Apr 2 15:42:05 1998
High retransmissions, check TCP patch level
    tcpIn tcpOut InConn/s OutConn/s %Retran ListenDrop/s Reset/s Attempt/s
      354 2714 0.00 0.37 32.17 0.00 0.00 0.00

But I only had a 1.2% collision rate as reported by netstat -i

Thanks to:
  Jim Robertori
  Miroslav Vrba
  Auteria Wally Winzer Jr.

Basically I was told that Adrian's default parameters are very low and what
I was seeing wasn't really a problem at all.

No one knew what "check TCP patch level" referred to.

However, this weekend I was upgrading a HPUX 10.20 server and added it to
the FDDI backplane of the switch, and I noticed what appeared to be network
loop between the LAN and FDDI interfaces of this HP. The switch went mad and
locked all segments solid. Since we already had one of these boxes in this
config, it is very possible that we had a network loop somewhere...
Broadcast storm I would guess...
Whether this would cause "high retransmissions" or not, I don't really know.

rgds
Stephen



This archive was generated by hypermail 2.1.2 : Fri Sep 28 2001 - 23:12:35 CDT