TCP/IP Internet interface error - temporary error - hosts

[Peter Barkas]

Since installing the latest TCP/IP ECO V5.4 for Alpha OpenVMS 7.3-2 I have been experiencing problem with host names, please see FTP example below, similar problem with RSH.

Anyone have any idea why?

[Steven Schweda]

Rather than saying "latest", show the output
from "tcpip show version".

Also, you might get more help sooner if your
posting actually showed the problem.

[Peter Barkas]

Original posting did include further information, but this web site was repeatedly failing and running very very slowly and I didn't even realise that the post had eventually worked.

$ tcpip show version

HP TCP/IP Services for OpenVMS Alpha Version V5.4 - ECO 6
on a AlphaServer DS10 617 MHz running OpenVMS V7.3-2

FTP> open fileserv1
%TCPIP-E-FTP_NETERR, I/O error on network device
-TCPIP-W-EAI_AGAIN, temporary failure

$ tcpip sho host fileserv1

LOCAL database

Host address Host name

172.16.1.19 FILESERV1, fileserv1, FILESERV2, fileserv2, FILESERV5,
fileserv5

FTP> open 172.16.1.19
220 Microsoft FTP Service
Connected to 172.16.1.19.
Name (172.16.1.19:barkas):

[Steve-Thompson]

Good morning Peter

Obviously need more info on this one...

I would use TCPIP$CONFIG and go through the interface configuration and check the state of the required services.

Then, enter into TCPIP> and do
sho interf /fu
sho service

and check the results.

Does Telnet work, in or out?
Does ping to an external node work?

Was anything else changed at the time of this upgrade, eg. the ethernet switch and/or cable?

Use $MC LANCP sho dev eia0 /char
Check line speed, media and Duplex???

Good luck

[Peter Barkas]

Thanks for the response.

Everything works fine, including FTP and TELNET and SSH both in and out, and seems normal if I use IP addresses.

However, any attempt to use a host name on an outgoing connection fails with the error shown.

As far as I am aware the only system level change was installing the ECO and rebooting.

[Steve-Thompson]

HAve you tried ...
1
Add a NEW entry to the host file, say a pc, and do a ping to the name.

2
How much of a job is replacing the "host file".

A question.
Is the service BIND in use on the VMS box?

[Peter Barkas]

Same results with a new entry to the hosts file.

Tried renaming the host file and running tcpip$config (which created a new file) and adding hosts, same result.

Bind is not configured.

Trace route reports:

"The name could not be resolved this time. Future attempts may succeed."

[Chinraj Rajasekaran]

Peter,

Try reloading the BIND database using

$ tcpip set name /init

otherwise check if bind resolved enabled

$ tcpip show name


regards
Raj

[Peter Barkas]

Raj

Thanks for the response.

As I mentioned above, Bind is not configured.

$ tcpip set name/init
rndc: neither TCPIP$ETC:RNDC.CONF nor TCPIP$ETC:RNDC.KEY was found. Use
'$rndc_confgen -a' to generate rndc.key or '$rndc_confgen' to generate rndc.conf

$ tcpip sho name

BIND Resolver Parameters

Local domain: co.uk

System

State: Started, Disabled

Transport: UDP
Domain:
Retry: Not defined
Timeout: Not defined
Servers: No values defined
Path: No values defined

Process

State: Disabled

[Steve-Thompson]

Peter

BIND is not in use which is fine! I would leave it that way!

It's an interesting problem....

1
Check the file security on tcpip$hosts.dat
Owner is SYSTEM with W:RE protection.
2
Dare I ask, did you do a reboot after the patch?

[Chinraj Rajasekaran]

Peter,

enable bind resolver

$ tcpip set name /enable /system

regards
Raj

[Peter Barkas]

Protection is w:re and owner is system.

The system was rebooted after the installing the patch, but not immediately, in fact it appears that it may have been a few days later.

[Chinraj Rajasekaran]

Peter,

Sorry..!

The BIND resolver should be diabled.

I see both system and process level status is disabled.

Then all name and address lookups
are now directed to the local hosts database.

So current settings are fine.

regards
Raj

[Steve-Thompson]

Hi Peter
ANy news?

I installed a 7.3-2 on a spare system and patched TCP/IP .... Sorry, HOSTS worked, before and after the reboot!

What more???
Can you go back to a previous OS backup image?

Do a $PROD SHO HIST and revise the installed patches.
Try RE-installing the TCPIP Patch. Shutdown TCPIP first. (Although I did it with TCPIP up "just for fun", I wouldnt do that normally)

Do a $prod remove of the package ...

Sorry, just some ideas!!!
Over to you!

[Peter Barkas]

Thanks all day every day.

I'm busy on a project just now, so I'll struggle by with it as it is, it's a pain but I can use IP addresses for now.

When I have time I will try your suggestions.

[Stefan Risch]

Hi,
i have similar problems with telnet-client and ftp-client on OpenVMS 7.3-2 with TCPIPV5.4 Eco 6.
The problem is in the ECO 6. I created a new system disk without further layered products or application software to prevent posible side-effects. Then i installed all rating-1 patches. Things worked fine until i installed the ECO 6....

Stefan Risch

[Peter Barkas]

Stefan, thank you for confirming that the problem is with ECO 6.

[Stefan Risch]

Here some additional informations about the testing environment:
$ product show hist
----------------------------------- ----------- ----------- --------------------
PRODUCT KIT TYPE OPERATION DATE AND TIME
----------------------------------- ----------- ----------- --------------------
DEC AXPVMS TCPIP_ECO V5.4-156 Patch Install 25-SEP-2006 16:43:17
DEC AXPVMS TCPIP_ECO V5.4-155 Patch Install 25-SEP-2006 15:04:04
DEC AXPVMS VMS732_LOADSS V1.0 Patch Install 25-SEP-2006 14:58:12
DEC AXPVMS VMS732_F11X V5.0 Patch Install 25-SEP-2006 13:45:52
DEC AXPVMS VMS732_DDTM V1.0 Patch Install 25-SEP-2006 13:41:26
DEC AXPVMS VMS732_SYS V10.0 Patch Install 25-SEP-2006 12:15:46
DEC AXPVMS VMS732_LMF V2.0 Patch Install 25-SEP-2006 12:10:55
DEC AXPVMS VMS732_ACRTL V2.0 Patch Install 25-SEP-2006 11:43:35
DEC AXPVMS VMS732_INSTAL V2.0 Patch Install 25-SEP-2006 09:01:46
DEC AXPVMS VMS732_UPDATE V7.0 Patch Install 25-SEP-2006 08:57:01
DEC AXPVMS VMS732_PCSI V3.0 Patch Install 25-SEP-2006 08:47:00
CPQ AXPVMS CDSA V2.0-109 Full LP Install 22-SEP-2006 16:45:58
DEC AXPVMS DECNET_PHASE_IV V7.3-2 Full LP Install 22-SEP-2006 16:45:58
DEC AXPVMS DWMOTIF V1.3-1 Full LP Install 22-SEP-2006 16:45:58
DEC AXPVMS OPENVMS V7.3-2 Platform Install 22-SEP-2006 16:45:58
DEC AXPVMS TCPIP V5.4-15 Full LP Install 22-SEP-2006 16:45:58
DEC AXPVMS VMS V7.3-2 Oper System Install 22-SEP-2006 16:45:58
HP AXPVMS KERBEROS V2.0-6 Full LP Install 22-SEP-2006 16:45:58
----------------------------------- ----------- ----------- --------------------

18 items found
$ tcpip show host

LOCAL database

Host address Host name

127.0.0.1 LOCALHOST, localhost
10.146.1.10 ldtest, LDTEST
$ telnet 10.146.1.10
%TELNET-E-IVHOST, Invalid or unknown host 10.146.1.10
-TCPIP-W-EAI_AGAIN, temporary failure
$ telnet ldtest
%TELNET-E-IVHOST, Invalid or unknown host ldtest
-TCPIP-W-EAI_AGAIN, temporary failure
$

[Wim Van den Wyngaert]

Always let patches ripe for a few months before installing them.

http://www4.itrc.hp.com/service/patch/patchDetail.do?patchid=DEC-AXPVMS-TCPIP_ECO-V0504-156-4&sel={openvms:alpha:7.3-2,}&BC=main|search| is not even showing the release notes. You need to download the patch first. Bad and sad.

Wim

[Jan van den Ende]

Peter, Stefan,

Thanks for the warning!
Saves us from getting hurt also.

Has Engeneering already be informed?
I would assume the patch to be put on hold ASAP, and a corrected one issued.

Proost.

Have one on me.

jpe

[Peter Barkas]

Jan

It seems that not all systems are affected, for example, 24x7 was unable to reproduce the problem, see earlier posting.

So until Stefan confirmed that he had reproduced the problem, I thought that it was an issue with my system.

So I haven't attempted to escalate this.

[Jan van den Ende]

So,

now it would be interesting to find out what is the same for you and Stefan, yet different for 24X7.
Might be non-trivial, but nevertheless, there obviously _DO_ exist (combinations of) circumstances that make this ECO non-generic.
And I do think, that ECOs should work for all configs that MIGHT be found "in the wild", or at least have their domain of valid configs explicitly stated.

After all, this _IS_ VMS, and such trustworthyness _IS_ what VMS stands for!

Proost.

Have one on me.

jpe

[Peter Barkas]

Anecdotal evidence mounting.

http://forums1.itrc.hp.com/service/forums/questionanswer.do?threadId=1062375

[STEPHEN CONWAY_1]

This is a regression in the bind resolver code. The problem could be avoided if one used the $IPC_SHR.EXE image from ECO5. Of course that is an unsupported thing to do.

Escalate for the fix. I don't know how to directly advertise fixes...

Steve

[Peter Barkas]

I extracted the image named above from ECO5, and installed it. Unfortunately, the problem persists.