T4 does not collect data

[Paul Jerrom]

Howdy,

I have 2xRX2620s clustered, running VMS V8.3. I have T4 installed on both, running the same command files, executables etc, logging to the same directories. One node runs fine; the other doesn't!

I can recreate the error - below is hacked from the T4$COLLECT command file:

$ Define Monitor T4$Sys:Monitor_IA64.Exe
$ monitor -
/Record=T4_LOKI_21JUN2007_1415_2359_Mon.Dat -
/Interval=60 -
/Flush_Interval=60 -
/Begin="21-JUN-2007 15:07:00.00" -
/End="21-JUN-2007 23:59:00.00" -
/Comment="HP rx2620 (1.60GHz/3.0MB) (2047Mb with 1 cpu(s))"-
/NoDisplay All_Classes
%MONITOR-I-NOUSER, TOPUSER data is not available in the requested input file

%MONITOR-E-COLLERR, error during data collection
-SYSTEM-F-BADPARAM, bad parameter value

...and I get no data collected. I've tried from various accounts, created a super account with more quota than necessary etc etc. Any ideas? Anyone aware of the parameter that might be bad?

Thanks
PJ

[Jon Pinkley]

Are both nodes using the same system disk? If so, I am surprised that one would work and the other would not. If each has it own system disk, does one system disk have new patches/updates (if so this is probably the one that does not work).

Try commenting out the "$ DEFINE MONITOR ..."
and use standard monitor. See if that caused the error to go away. The samples collected may drift by a few seconds over the day, but other than that it should work.

[John Gillings]

PJ,
I don't have an equivalent system to test on. That TOPUSER error sounds suspicious. Is that a new in V8.3 thing? Are you using the latest T4 kit?


Maybe try with the normal MONITOR image. Do you get the same error?

Try the T4 image with just "MONITOR ALL" what happens?

If that doesn't help, I recommend you log a case with your local customer support centre, and/or send an email to the t4 comments address t4hpcom, or use the T4 comment page http://h71000.www7.hp.com/openvms/products/t4/fb_t4mailto.html

[Willem Grooters]

Be sure to have the latest version of the kit.
Next is a requirement for Alpha because there are some problems with the specific monitor version, but I've understood it applies to Itanium as well:

In t4$COLLECT.COM, look for this part of code:

$ If (Version_For_Compare .Le. 732)
$ Then
$ Define Monitor_Tv T4$Sys:Monitor_Tv_No_Drift.exe
$ Else
$! USE NORMAL MONITOR
$ If (Arch_Name .eqs. "ALPHA") Then Define Monitor T4$Sys:Monitor_Alpha.Exe
$ If (Arch_Name .eqs. "IA64") Then Define Monitor T4$Sys:Monitor_IA64.Exe
$ EndIf

Comment the definition of the Monitor logical:

$!! If (Arch_Name .eqs. "IA64") Then Define Monitor T4$Sys:Monitor_IA64.Exe

It seems to be 8.3 specific, HP is aware of the problem.

Willem

[Willem Grooters]

Correction: The change in the commandprocedure is NOT 8.3 specific, but this is: Extraction of T4-data from monitor output will give this error:

%SYSTEM-W-BADFILEHDR, bad file header

for each extraction. This will lead to numerous "Fine not found" errors later on.

This has been reported to HP, but I'm not sure what the current state is.

WG

[Volker Halle]

Paul,

just to confirm: I have downloaded and installed HP-VMS-T4-V0400--1 on a brand-new OpenVMS I64 rx2620 cluster (all patches applied) 2 weeks ago and T4 worked without a problem...

Volker.

[Paul Jerrom]

Hi all,

Thanks for the interest. Version I am using is HP-VMS-T4-V0400--1. Both nodes use the same t4$sys directory (on a volume shadowed, shared logical disk). AFAIK the two servers are installed & patched identically - I tend to do things to both servers at the same time.
Actually, the TOPUSER error occurs on both nodes, but on the other node the data collection is fine.
I will try without the redefinition of the monitor logical.
All the best,
PJ

[Volker Halle]

Paul,

after checking the T4$DATA:T4_node_26JUN2007_0001_2359_SUBP_MON.LOG file, I must admit, that I'm also seeing the following message on both nodes:

%MONITOR-I-NOUSER, TOPUSER data is not available in the requested input file

This error message is a little bit misleading, as it is issued for a MONITOR/RECORD operation, where there is no 'input' file.

But this can be explained very easily: the MONITOR_I64.EXE image (built 4-FEB-2005) delivered with T4 V4.0 is not based on V8.3 and therefore does NOT contain the new (V8.3) /TOPUSER qualifier.

Does the %MONITOR-E-COLLERR/BADPARAM problem happen immediately or some time after starting this monitor session ?

Volker.

[Volker Halle]

Paul,

I stronlgy believe that the

%MONITOR-E-COLLERR, error during data collection
-SYSTEM-F-BADPARAM, bad parameter value

error is independent of the MONITOR-I-NOUSER informational message.

This error will most likely be related to some configuration/system parameter combination. Only HP - with access to the sources of MONITOR_IA64.EXE - can find out what's wrong.

Volker.

[Paul Jerrom]

Howdy,

Yes, the two MONITOR errors are independant of each other - both nodes get the first error, but only one node gets the second.

The BADPARAM error occurs immediately, and not after a delay.

Have tried using standard MONITOR instead of that supplied with T4, but this causes a "%SYSTEM-W-BADFILEHDR, bad file header" error when T4EXTR tries to process the files - so I end up with no data for either node!

T4 ain't gonna be my friend for much longer...