Messages in this thread | | | From | David Woodhouse <> | Subject | TCP stuck (RH 2.2.12-20 <-> 2.2.13) | Date | Sat, 19 Feb 2000 01:49:35 +0000 |
| |
I've got a TCP connexion stuck. It's a SSH connexion from gate.mvhi.com to lnx.keraben.com. The link is generally fairly appalling, but at this time of night it's not so bad,
netstat on lnx.keraben.com says: tcp 0 37956 lnx.keraben.com:ssh gate.mvhi.com:1021 ESTABLISHED
netstat on gate.mvhi.com says: tcp 0 0 gate.mvhi.com:1021 lnx.keraben.com:22 ESTABLISHED
Typing into my end of the session gives immediate acks:
tcpdump taken at gate.mvhi.com:
01:39:01.250977 gate.mvhi.com.1021 > lnx.keraben.com.22: P 40:60(20) ack 1 win 31856 <nop,nop,timestamp 766234177 97749147> (DF) [tos 0x10] 01:39:01.392829 lnx.keraben.com.22 > gate.mvhi.com.1021: . ack 60 win 31856 <nop,nop,timestamp 97750075 766234177> (DF) [tos 0x10]
tcpdump taken at lnx.keraben.com:
02:39:01.279777 < gate.mvhi.com.1021 > lnx.keraben.com.ssh: P 40:60(20) ack 1 win 31856 <nop,nop,timestamp 766234177 97749147> (DF) [tos 0x10] 02:39:01.293987 > lnx.keraben.com.ssh > gate.mvhi.com.1021: . 1:1(0) ack 60 win 31856 <nop,nop,timestamp 97750075 766234177> (DF) [tos 0x10]
However, the remote machine (lnx.keraben.com) hasn't attempted to clear it's send-q for at least half an hour.
The remote machine is Red Hat's 2.2.12-20 kernel, and the machine at my end is running 2.2.13.
Other connections to the same machine are working fine.
I don't know what triggered it to get stuck. I was running an X client over ssh over the link - but I've done that quite a lot recently, and although it's fairly slow, it's not utterly insane. Whatever happens, it ought not get stuck.
NOTE: I've been running a tcpdump at the remote end of the link too - it's not just that the data packets are getting dropped in transit. They're really not being sent.
-- dwmw2
- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.rutgers.edu Please read the FAQ at http://www.tux.org/lkml/
| |