Error: could not resolve start-time

[jose_ferrada]

Hello I run vmmark for first time (after any days working in setup and correct errors), in the last run I execute TURBOMODE ON for tile 0 and WORKLOADLIST="DS2WebA DS2WebB DS2WebC"  after vmmark end process this error appear in the score.out

VMmark 2.5 TileScore : v1.5 : updated 01292014

Computing_results_for_test_with_tile_count: 0 ...

Turbo Mode Enabled

Tiles = 1 : Enabled Workloads : DSDWeb  (3)

Use of uninitialized value in concatenation (.) or string at C:\vmmark2\tilescore.pl line 2192.

First Sample: 1475372280 Sun Oct  2 01:38:00 2016

Info: 29 :  : 1 : 3

Undefined times at 29

Calculating Turbo Timing & Scoring : Run_Is_NOT_Compliant

Run_start 1475372280  : Sun Oct  2 01:38:00 2016

Start_time 0 : Thu Jan  1 00:00:00 1970

End_time 1475373960 : Sun Oct  2 02:06:00 2016

Run_end 1475373960 : Sun Oct  2 02:06:00 2016

Duration_in_minutes : 24589566.00

Steady_start_start 300 : Thu Jan  1 00:05:00 1970

Steady_state_end 1200 : Thu Jan  1 00:20:00 1970

Phase_0_begin 300 : Thu Jan  1 00:05:00 1970

Phase_1_begin 600 : Thu Jan  1 00:10:00 1970

Phase_2_begin 900 : Thu Jan  1 00:15:00 1970

Error: could not resolve start-time. Results data missing from one or more *.wrf files or time on clients not synchronized

[RebeccaG]

Hi Jose, I apologize for my delay in responding. There was an issue with the notifications for this VMmark community.

First, all the VMmark workloads have to be enabled for Score_0_Tile_Test.out to run correctly.

Second, your DS2 Web VMs are not running correctly because you did not enable the DS2DB VM.

Change

WORKLOADLIST="DS2WebA DS2WebB DS2WebC" to

WORKLOADLIST="DS2WebA DS2WebB DS2WebC DS2DB"

and rerun.

You can post your run back on this thread if you would like us to take a look at it and help you get going.

[jose_ferrada]

hi Rebecca thanks for your reply.

I followed your recommendations and I am now running all workloads with TurboMode on (30 minutes) and generates the Score_Turbo_1_Tile_Test.out file. I have not enabled vmotion since the vmware cluster is only one node.

but when you run the test by default (3 hours) the Score file is not generated, I have not enabled vmotion since the vmware cluster is only one node.

when it is running in active mode TurboMode the result of the score as it should be compared with the values indicating the documentation?

Table 1-7. Reference System Workload Scores

Workload Name Score

Standby server None

Mail server 330.25 send mail actions/minute

Olio 4641.43 operations/minute

DS2WebA 2199.18 operations/minute

DS2WebB 1518.55 operations/minute

DS2WebC 1058.05 operations/minute

Clone and deploy 4.0 DeployedVMs/Hour

vMotion 16.0 vMotionedVMs/Hour

Storage vMotion 9.0 Storage vMotionedVMs/Hour

[jose_ferrada]

This is the result of execution for 30 minutes

[jose_ferrada]

This is the result of execution for 3 hours

[RebeccaG]

Hi Jose, vMotion is also a workload so that's why the 3 hour test is not generating Score.out.

Typically VMmark is run on a 2-host cluster, which is required to produce a VMmark score result like "1.35".

However, you can create a Score.out manually for the 3 hour test where you can look at the scores for all the other workloads that are running.

Look in the VMmark Benchmarking Guide, under "Obtain Partial Results for a Failed or Non-compliant Run".

The command you will want to run is

perl ../../tilescore.pl -t 1 -V

when it is running in active mode TurboMode the result of the score as it should be compared with the values indicating the documentation?

Results from turbo mode should not be compared to the numbers in the documentation. But for a 3 hour test, yes, you can compare the numbers in the documentation to your numbers to see if VMmark performance is in the right 'ballpark'.

In Score.out, look for "Tile_0_Ratios:" The numbers in this block should be >= 1.0.

But the values in the documentation are just an example of how VMmark might run. Your numbers will be different and cannot be directly compared. Usually, people who run VMmark run it in different environments or with different configurations, and then compare their results to each other.

[jose_ferrada]

Hi Rebecca thanks for your reply.

I am grateful for the dedication and clarity in its response.

It was invaluable your help to understand and resolve the issues presented in the execution of VMmark.

have a great day.