Message boards : ATLAS application : Status: Waiting for memory (5 CPUs) - finished but won't upload
Message board moderation

To post messages, you must log in.

AuthorMessage
Bradders

Send message
Joined: 3 Jan 17
Posts: 13
Credit: 497,855
RAC: 0
Message 30435 - Posted: 20 May 2017, 8:08:50 UTC

I have run an ALTAS Simulation 1.01 (vbox64_mt_mcore_atlas) WU on my dual Xeon CPU quad core PC. E5345 @ 2.33GHz with 8GB RAM.

The WU progress is 100%, Elapsed time is 12d 19:36:09 and Remaining time reads --- and the deadline is 20/05/2017 2:49:25.

However, the WU won't upload.
The Status reads "Waiting for memory (5 CPUs)".
Any ideas how to make it upload?
ID: 30435 · Report as offensive     Reply Quote
maeax

Send message
Joined: 2 May 07
Posts: 2071
Credit: 156,086,835
RAC: 104,225
Message 30436 - Posted: 20 May 2017, 8:39:37 UTC - in response to Message 30435.  
Last modified: 20 May 2017, 8:41:01 UTC

in Boinc-manager under Transfers - retry now for this task.

Otherwise save Boinc-manager and wait until Virtualbox have saved the boinc-task.

Then restart Boinc-manager again.
ID: 30436 · Report as offensive     Reply Quote
computezrmle
Volunteer moderator
Volunteer developer
Volunteer tester
Help desk expert
Avatar

Send message
Joined: 15 Jun 08
Posts: 2386
Credit: 222,904,781
RAC: 137,963
Message 30437 - Posted: 20 May 2017, 8:43:48 UTC - in response to Message 30435.  

I have run an ALTAS Simulation 1.01 (vbox64_mt_mcore_atlas) WU on my dual Xeon CPU quad core PC. E5345 @ 2.33GHz with 8GB RAM.

The WU progress is 100%, Elapsed time is 12d 19:36:09 and Remaining time reads --- and the deadline is 20/05/2017 2:49:25.

However, the WU won't upload.
The Status reads "Waiting for memory (5 CPUs)".
Any ideas how to make it upload?

A 5-core WU requests 6.6 GB RAM.
This is >80% on your 8 GB host.
You may check if your BOINC client is allowed to use more than 80 % RAM.

The error will reoccure if your client runs WUs (or keeps paused WUs in RAM) with a total RAM requirement that exceeds your preferences.
ID: 30437 · Report as offensive     Reply Quote
Bradders

Send message
Joined: 3 Jan 17
Posts: 13
Credit: 497,855
RAC: 0
Message 30438 - Posted: 20 May 2017, 9:15:33 UTC - in response to Message 30436.  

in Boinc-manager under Transfers - retry now for this task.
Otherwise save Boinc-manager and wait until Virtualbox have saved the boinc-task.
Then restart Boinc-manager again.

I should have explained that it was not in the Transfers list.
I closed BOINC, opened VBox Manager v5.1.12 and re-opened BOINC. The WU now has status Running (5 CPUs).
The elapsed time is running, but the WU is still at 100.000%
ID: 30438 · Report as offensive     Reply Quote
Bradders

Send message
Joined: 3 Jan 17
Posts: 13
Credit: 497,855
RAC: 0
Message 30439 - Posted: 20 May 2017, 9:17:15 UTC - in response to Message 30437.  

I have run an ALTAS Simulation 1.01 (vbox64_mt_mcore_atlas) WU on my dual Xeon CPU quad core PC. E5345 @ 2.33GHz with 8GB RAM.

The WU progress is 100%, Elapsed time is 12d 19:36:09 and Remaining time reads --- and the deadline is 20/05/2017 2:49:25.

However, the WU won't upload.
The Status reads "Waiting for memory (5 CPUs)".
Any ideas how to make it upload?

A 5-core WU requests 6.6 GB RAM.
This is >80% on your 8 GB host.
You may check if your BOINC client is allowed to use more than 80 % RAM.

The error will reoccur if your client runs WUs (or keeps paused WUs in RAM) with a total RAM requirement that exceeds your preferences.

I changed the memory limit to 85%.

Sadly, it looks like the job aborted.
ID: 30439 · Report as offensive     Reply Quote
Bradders

Send message
Joined: 3 Jan 17
Posts: 13
Credit: 497,855
RAC: 0
Message 30440 - Posted: 20 May 2017, 9:22:00 UTC

I have a new 5 CPU WU. It estimates about 3 hours, not 12 days!
I should leave the PC unattended for 2 weeks!
ID: 30440 · Report as offensive     Reply Quote
maeax

Send message
Joined: 2 May 07
Posts: 2071
Credit: 156,086,835
RAC: 104,225
Message 30441 - Posted: 20 May 2017, 9:24:10 UTC

ID: 30441 · Report as offensive     Reply Quote
computezrmle
Volunteer moderator
Volunteer developer
Volunteer tester
Help desk expert
Avatar

Send message
Joined: 15 Jun 08
Posts: 2386
Credit: 222,904,781
RAC: 137,963
Message 30442 - Posted: 20 May 2017, 12:55:04 UTC

@Bradders
You may, as already mentioned, work through Yeti's checklist.

Beside that you may start from the scratch and try to successfully finish a couple of Theory WUs as this is the CERN-VM-subproject with the least requirements.
Set Max# jobs to 1 and Max# CPUs also to 1 on the project's website.
Once you are familiar with Theory go ahead with other subprojects.
Always start with a single 1-core task before running more tasks concurrently.

The following (simple) app_config.xml may also help.
Place it in your folder "Path\projects\lhcathome.cern.ch_lhcathome" and reload your client's configuration.
This saves your already downloaded WUs (except that ones that are already started).

<app_config>
  <app>
    <name>ATLAS</name>
    <max_concurrent>1</max_concurrent>
    <fraction_done_exact/>
  </app>
  <app_version>
    <app_name>ATLAS</app_name>
    <plan_class>vbox64_mt_mcore_atlas</plan_class>
    <avg_ncpus>2.0</avg_ncpus>
    <cmdline>--nthreads 2 --memory_size_mb 4600</cmdline>
  </app_version>
  <app>
    <name>CMS</name>
    <max_concurrent>1</max_concurrent>
  </app>
  <app>
    <name>LHCb</name>
    <max_concurrent>1</max_concurrent>
  </app>
  <app>
    <name>Theory</name>
    <max_concurrent>1</max_concurrent>
  </app>
  <project_max_concurrent>1</project_max_concurrent>
</app_config>
ID: 30442 · Report as offensive     Reply Quote
Bradders

Send message
Joined: 3 Jan 17
Posts: 13
Credit: 497,855
RAC: 0
Message 30485 - Posted: 24 May 2017, 22:53:30 UTC - in response to Message 30441.  

This is the checklist from Yeti:

https://lhcathome.cern.ch/lhcathome/forum_thread.php?id=4161#29359

Thanks
Here is what I found going through the checklist:
4. Virtualbox is 5.1.12, which is behind the 5.1.16 recommendation. 5.1.22 downloaded.
6. I have 8 cores and 8GB of memory.
I changed my settings to use a maximum of 2 CPUs, aborted the 5CPU WU and updated BOINC.
14. Yeah, there is a big difference between 12d and 4h!

I have kept ATLAS on the list of WU, but I won't be running a 5CPU job unless I double the RAM.

Thanks again everyone.
ID: 30485 · Report as offensive     Reply Quote

Message boards : ATLAS application : Status: Waiting for memory (5 CPUs) - finished but won't upload


©2024 CERN