There appears to be a bug in rsh, and possibly rcp, in the RH 6.2
distribution both on Intel and Alpha.
Our VMS box has problems talking to the RH 6.2 machines (Alpha) with rsh,
rcp, etc. I'm also seeing really poor rcp performance between two
Linux/Alpha machines (DS10). rcp moves files very slowly, while
ftp/NFS copies move the same files much more quickly. I suspect that
the two problems may be related as both rsh and rcp come from the same
RPM.
The problem with rsh is that sometimes these work, and sometimes they
hang and then time out. That is, if I do:
$ rsh linuxbox "ls"
several times it will eventually hang and timeout on one of them. The
vendor of the TCP/IP stack for VMS swears that it's a bug on the Linux
side and can support that with TCPDUMP. They observe that their side
(the VMS side) sends a SYN but the Unix side never ACKs. (See below)
Is this a known bug? Can the person who's in charge of rsh/rcp etc.
please have a look at it? The affected rsh distribution is
"rsh-0.16-12" (and I've seen it on earlier ones as well).
Here is TCPDUMP for the transaction, recorded by Hunter Goatley
([EMAIL PROTECTED])
From: Hunter Goatley ([EMAIL PROTECTED])
Subject: RE: rcp, rsh variants that don't hang?
Newsgroups: vmsnet.networks.tcp-ip.multinet
Date: 2000-08-31 10:26:32 PST
Hunter Goatley <[EMAIL PROTECTED]> writes:
>
>Seeing a TCPDUMP showing the exchange (those that work and those that
>don't) would be most helpful. I also plan to install Linux here
>shortly, which will give me a chance to try to reproduce this myself.
>
OK, here's part of a TCPDUMP we just created using RedHat Linux V6.2
and MultiNet V4.3. At 13:05, you see the ending a successful RSH,
then at 13:06:01.01, you see the next one. We send out 4 SYNs, the
other side never ACKs, it eventually times out.
13:05:59.65 yyz.process.com.1023 > zzy.process.com.shell: F 29:29(0) ack 591 win 6144
<,,timestamp 458752:-1281753088> (DF)
13:05:59.65 yyz.process.com.1022 > zzy.process.com.1023: F 1:1(0) ack 2 win 6144
<,,timestamp 458752:-1281753088> (DF)
13:05:59.65 zzy.process.com.shell > yyz.process.com.1023: . ack 30 win 32120
<,,timestamp 7667712:1838546944> (DF)
13:05:59.65 zzy.process.com.1023 > yyz.process.com.1022: . ack 2 win 32120
<,,timestamp 7667712:1838546944> (DF)
13:06:01.01 yyz.process.com.1023 > zzy.process.com.shell: S 0:0(0) win 6144 <mss
1460,,wscale 0,,,timestamp 458752:-1281556480> (DF)
13:06:06.75 yyz.process.com.1023 > zzy.process.com.shell: S 0:0(0) win 6144 <mss
1460,,wscale 0,,,timestamp 458752:-1280835584> (DF)
13:06:18.77 yyz.process.com.1023 > zzy.process.com.shell: S 0:0(0) win 6144 <mss
1460,,wscale 0,,,timestamp 458752:-1279262720> (DF)
13:06:42.82 yyz.process.com.1023 > zzy.process.com.shell: S 0:0(0) win 6144 <mss
1460,,wscale 0,,,timestamp 458752:-1276116992> (DF)
Thanks,
David Mathog
[EMAIL PROTECTED]
Manager, sequence analysis facility, biology division, Caltech
_______________________________________________
Redhat-devel-list mailing list
[EMAIL PROTECTED]
https://listman.redhat.com/mailman/listinfo/redhat-devel-list
