DTSS$CLERK doesn't synchronize the time

[Thomas Schick]

My environment is OpenVMS V7.3 / DECnet-Plus ECO03 / TCPIP V5.1 ECO04.
DTSS$CLERK doesn't work; "Status state" remains in "Synchronizing" and
$MCR NCL SHOW DTSS DECNET GLOBAL SERVER *
brings
command failed due to:
no such object instance
In the same configuration as in DECnet-Plus ECO03, the ECO02 works well.

What kind of problem does the output in DTSS$ERROR.LOG signal?
E GetBindData: gethostbyname error: name: gbl_dtss1, errno: 50, vaxc$errno: 8340
E GetBindData: gethostbyname errno: network is down
E GetBindData: gethostbyname h_errno: Error 0
E PopulateCacheFromBind:Call GetBindData Failed, sysStatus=2
:
:
Any idea?

Cheers, Thomas

[Wim De Cock]

Hello,

code 50 means %SYSTEM-E-BADATTRIB, bad attribute control list
code 8340 means %SYSTEM-F-UNREACHABLE, remote node is not currently reachable

--> It looks like the name translation fails (and returns an empty string or so) and thus the connect fails.
Does "$mc tcpip$nslookup gbl_dtss1" work ?
If not, this migth be the first issue to resolve.

If it works --> the code 50 would suggest a corrupted DTSS config file

Hope this helps ...

[John Eerenberg]

I can tell you that you can ignore the error message from your "ncl show dtss" command. A correctly configured client and clerk is all you need.

The clerk is an intermediate between the client and server.

The clerk errors could be because you still need to point to a time provider (either on the internet or one within your network). Have you declared a time provider to your system?

If you have no time provider and simply point a server or multiple servers, then your clock will track to the average all the systems and their clocks (that's over simplified but the idea is to minimize time drift).

I'll see if I can find the exact commands to point to a time provider or maybe someone has them handy. You'll want multiple servers and with access to time providers.

[Wim Van den Wyngaert]

Have the same problem today. But it ended with :

Fatal error at line 3227 in file ri_axp_eco_resd$:[dtss_service.src]dtss_decnet_v_vms.c
%LIB-F-BADLOADR, bad block address

Same VMS and Decnet version ad Thomas. No nameserving and no decnet db update at the moment of the crash.

The DTSS process aborted and I restarted decnet (via a reboot).

Probably caused by network interrupt.

Wim

[Heinz W Genhart]

Hi Thomas

We are using DECnet over IP (RFC1006) without any DECnet routing.
On those machines not belonging to same IP subnet as the DTSS Server, we defined the address of the global server in our local TCP/IP database as follows:

$ TCPIP SHO HOST /local

1.2.3.4 GBL_DTSS1, gbl_dtss1
1.2.3.5 GBL_DTSS2, gbl_dtss2
1.2.3.4 GBL_DTSS3, gbl_dtss3

This hosts (gbl_dtss1 .. gbl_dtss3) will be used for synchronisation.

See also http://h71000.www7.hp.com/doc/82final/decnetplus/DTS_MNG.pdf chapter 2.2.4.2

Hope that helps

Regards

Heinz