ftp daemon takes 50 seconds to response

[eric_17]

Hi all,

Telnet and ping have no problem and very fast. But ftp daemon takes over 50 seconds to response. Any idea?

UCX> show service/full ftp

Service: FTP
State: Enabled
Port: 21 Protocol: TCP Address: 0.0.0.0
Inactivity: 5 User_name: UCX$FTP Process: UCX$FTPD
Limit: 10 Active: 0 Peak: 1

File: SYS$SYSDEVICE:[UCX$FTP]UCX$FTPD_STARTUP.COM
Flags: None

Socket Opts: Rcheck Scheck
Receive: 0 Send: 0

Log Opts: Acpt Actv Dactv Conn Error Exit Logi Logo Mdfy Rjct TimO Addr
File: SYS$SYSDEVICE:[UCX$FTP]UCX$FTPD_STARTUP.LOG

Security
Reject msg: not defined
Accept host: 0.0.0.0
Accept netw: 0.0.0.0

Thank you very much!

[labadie_1]

It would be interesting to know your Vms and Ucx/Tcpip version. You shoukd check you have the last Eco fro Ucx/ Tcpip.

This problem may be related to name resolution,

$ ucx sh name
and some others could help.

regards

Gerard

[Ian Miller.]

Could be dns reverse lookup problem - trying to look up IP address of incoming connection to get name.

Could be resource limitation slowing creation of process - memory, disk etc

[eric_17]

For sure there is no problem for name resove. nslookup works very well.

$ ucx sh name

BIND Resolver Parameters

Local domain: hcl.com

System

State: Started, Enabled

Transport: UDP
Domain: hcl.com
Retry: 4
Timeout: 4
Servers: x.x.x.x

Process

State: Enabled

Transport:
Domain:
Retry:
Timeout:
Servers:


Any idea? BTW, I am not good at DECVMS

Thanks

[labadie_1]

ucx sh version
sh sys/noproc

?

[eric_17]

$ ucx sh version

DEC TCP/IP Services for OpenVMS AXP Version V4.0 - ECO Level 2
on a AlphaStation 400 4/233 running OpenVMS V6.2

$ sh sys/noproc
OpenVMS V6.2 on node HCL1 9-MAR-2004 19:51:39.24 Uptime 82 23:20:10

[labadie_1]

You should put Ucx 4.2 Eco 5 and retry.

(and you will be on a supported version)

[eric_17]

Maybe I didn't make myself clear. From PC to DECVMS, telnet and ping have no prolbem, but FTP takes over 50 seconds.

C:\>ftp decvms
Connected to decvms.hcl.com.
Connection closed by remote host.

[labadie_1]

if you do from the Pc
ftp alpha_name

and ftp alpha-IPaddress

do you wait 50 seconds in both cases ?

[eric_17]

yes. It takes over 50 second on both IP and Machine name.

C:\>ftp decvms
Connected to decvms.hcl.com.
Connection closed by remote host.

C:\>ping decvms

Pinging decvms.hcl.com [10.1.88.244] with 32 bytes of data:

Reply from 10.1.88.244: bytes=32 time<10ms TTL=254

Ping statistics for 10.1.88.244:
Packets: Sent = 1, Received = 1, Lost = 0 (0% loss),
Approximate round trip times in milli-seconds:
Minimum = 0ms, Maximum = 0ms, Average = 0ms
Control-C
^C
C:\>ftp 10.1.88.244
Connected to 10.1.88.244.
Connection closed by remote host.

[labadie_1]

The Pc and Vms are on the same lan ?

from the PC:
tracert decvms

From Vms
$ trace :==$ ucx$examples:ucx$traceroute

$ trace PC

and anyway, go to ucx4.2 Eco 5 and reboot

[eric_17]

Yes, they are on the same subnet. How can I go to ucx4.2 Eco 5, do I need upgrade? If so, how?

[labadie_1]

You should ask your support center.

I think you will find the Eco 5 at ftp.itrc.hp.com

To install that, you have 2 possibilities, depending on the files you get

1) use
$ @sys$update:vmsinstal
if you have files .a, .b .c...

2) use
$ product instal */source=dka200:
if your file has a .pcsi extension, and you have put it in dka200:

Good luck

[labadie_1]

try on Vms
$ reply/ena

then from the Pc do your
ftp decvms

Do you have "interesting" messages ?

[eric_17]

$ reply/ena on the DECVMS, and ftp from the PC, I got "interesting" message


$
%%%%%%%%%%% OPCOM 9-MAR-2004 23:31:56.08 %%%%%%%%%%%
Message from user AUDIT$SERVER on HCL1
Security alarm (SECURITY) and security audit (SECURITY) on HCL1, system id: 10
Auditable event: Local interactive login failure
Event time: 9-MAR-2004 23:31:56.08
PID: 0000F65D
Process name: _TNA4231:
Username:
Process owner: [SYSTEM]
Terminal name: _TNA4231, _TNA4231:
Image name: HCL1$DKA0:[SYS0.SYSCOMMON.][SYSEXE]LOGINOUT.EXE
Status: %LOGIN-F-CMDINPUT, error reading command input

[Martin P.J. Zinser]

Hi,

and since it is not exactly trivial to locate something as old as UCX 4.0 ECO 5 on ITRC, here is the direct link:

ftp://ftp.itrc.hp.com/archived_patches/openvms_patches/axp/6.X/v6.1/ucx/4.0/

HTH,

Martin

[Wim Van den Wyngaert]

Eric,

Don't forget to re-qualify the applications before doing the upgrade.
I personally don't do any upgrades unless they are tested and approved by the application. Of course, if downtime is not important and applications are robust, you can always rollback. But that makes it a little more complicated ...

Wim

[Lokesh_2]

Hi,

Can you restart FTP and check whether it has resolved the problem ?

@sys$startup:ucx$ftpd_shutdown.com
@sys$startup:ucx$ftpd_startup.com

Best regards,
Lokesh

[Wim Van den Wyngaert]

Were you able to give a username ?

If yes :
Verify with another username if it is working. It may be the login.com doing something wrong. Or remove (rename) login.com so that it isn't executed.
In any case, restart ftp as said by Lokesh.

It could also be the sylogin.com ...

[Ian Miller.]

Could be a resource problem - do other process creations take a long time? or other users on same disk have problem?

Most likely is something in SYLOGIN or LOGIN so suggestions about using an different username are good.

In LOGIN.COM and SYLOGIN.COM you could do
$ SET VERIFY
$ SET PREFIX "(!5%T) "

to find out how long each step takes.

[eric_17]

Hi Lokesh, I have tried @sys$startup:ucx$ftpd_shutdown.com and @sys$startup:ucx$ftpd_startup.com. after restart the FTP daemon, it still doesn't work.

Van den Wyngaert Wim, how can I rename login.com file?

Thanks a lot!

[Lokesh_2]

Hi Eric,

Try to FTP from your VMS machine to itself.
Give the following command and check how long it takes:

$ftp localhost

To rename login file, set default to user home directory and create an empty login.com ( or whatever the lgicmd is set for user account ) with creat command.

HTH,
Best regards,
Lokesh Jain

[eric_17]

Hi Lokesh,

"ftp localhost" is very fast. I just re-checked the network parameters. It seems like something wrong for bind server. When I try to re-config a new bind Server name and IP, it shows me the machine is not in the local host database. How could I add a machine and IP to the local host database? Does vms system have a file which has same function as /etc/hosts file?

Regards,

[Lokesh_2]

Hi,

In vms the file name is SYS$SYSTEM:UCX$HOST.DAT . Anyway you cannot edit this file like etc/host in Unix.

To add a host name in local host database , give the following command:

$ucx set host /address=

To check the entries in local host database, give command:

$ucx sho host

HTH,
Best regards,
Lokesh Jain

[eric_17]

OK, "ucx show host" command shows that the machine and IP is in the database already. But it says Default servers are not available.