LHC@home

Atlas used to work on all my computers. I just upgraded two of them with extra RAM. They're dual Xeon machines which now have 36GB instead of 20GB of RAM. But I'm getting this: https://lhcathome.cern.ch/lhcathome/results.php?userid=55945&offset=0&show_names=0&state=5&appid=14 - all the tasks are only using 1 CPU core instead of 8, and finishing within 15 minutes, then causing a validate error. Any way to find out what's wrong? All I changed was adding more RAM (which has been tested by Memtest).

Looking at one of the task logs, I see this from https://lhcathome.cern.ch/lhcathome/result.php?resultid=275499071:

*****
2020-05-27 23:12:03 (3224): Guest Log: *** Error codes and diagnostics ***

2020-05-27 23:12:03 (3224): Guest Log: "exeErrorCode": 65,

2020-05-27 23:12:03 (3224): Guest Log: "exeErrorDiag": "Non-zero return code from EVNTtoHITS (33); Logfile error in log.EVNTtoHITS: \"DetectorStore FATAL in sysInitialize(): standard std::exception is caught\"",

2020-05-27 23:12:03 (3224): Guest Log: "pilotErrorCode": 1165,

2020-05-27 23:12:03 (3224): Guest Log: "pilotErrorDiag": "Local output file is missing",
*****

---

Same problem on my other computer which has not changed apart from upgrading Virtualbox and extensions to latest version.

I shall cease Atlas tasks until someone tells me what's happened.

---

... until someone tells me what's happened.

Just look around and read other posts.
The answer might already be there:
https://lhcathome.cern.ch/lhcathome/forum_thread.php?id=5438&postid=42630

As you can read in the above link, there was a major database outage at CERN yesterday evening which affected BOINC servers and pretty much all of ATLAS' distributed computing services. Unfortunately one of the last things to come back were the Frontier database servers which the ATLAS tasks read data from as they are running. So although we were able to submit tasks here, they would all fail straight away. Now things should be working ok, sorry for the inconvenience.

... until someone tells me what's happened.

Just look around and read other posts.
The answer might already be there:
https://lhcathome.cern.ch/lhcathome/forum_thread.php?id=5438&postid=42630

I did search first, but the search function in these forums isn't the best, and I was looking for a specific Atlas problem. It never dawned on me to link this problem to the outage I noticed when I couldn't even get on the forums yesterday.

As you can read in the above link, there was a major database outage at CERN yesterday evening which affected BOINC servers and pretty much all of ATLAS' distributed computing services. Unfortunately one of the last things to come back were the Frontier database servers which the ATLAS tasks read data from as they are running. So although we were able to submit tasks here, they would all fail straight away. Now things should be working ok, sorry for the inconvenience.

No problem, I just thought I was doing something wrong, I didn't want to throw back hundreds of useless results.

And now processing with 8 cores per Atlas task, so I assume all is ok.

---

During the night and this morning I've had quite a lot of invalid tasks (>25). All those tasks have failed with other hosts as well. The error for all of them seems to be 'Error: Service 'control' failed to initialize: VERR_INVALID_PARAMETER'.

Here's one of them: https://lhcathome.cern.ch/lhcathome/result.php?resultid=277054761

---

Since 2 hours i get nothing but validation errors on ATLAS native tasks!!

The problem seems to be:
"pilotErrorDiag": "Transform not found:/bin/bash: Sim_tf.py: command not found\n"

There are a lot of other hosts out there with failing tasks!

---

Why mine when you can research? - GRIDCOIN - Real cryptocurrency without wasting hashes! https://gridcoin.us

Now all my Atlas tasks (windows virtual box) are failing with the error I posted before. I put Atlas on hold for a while.

---

In addition to the validating errors on ATLAS i have now troubles getting other LHC workunits.

BOINC tells me:
Fr 12 Jun 2020 20:04:56 CEST | LHC@home | Scheduler request failed: HTTP gateway timeout

Uploading results seems to be fine.

---

Why mine when you can research? - GRIDCOIN - Real cryptocurrency without wasting hashes! https://gridcoin.us

In addition to the validating errors on ATLAS i have now troubles getting other LHC workunits.

I had trouble too until I read LHC BOINC Messages and saw that no CPU was requested because queue was full and none needed. I suspended unstarted WUs and LHC immediately DLed a boatload. Hopefully this batch won't fail instantly.

Edit: Not looking good: Valids zero, Invalids 73. Validation error.

Validate error on all my ATLAS tasks today.

same here, on 2 machines so far:

2020-06-13 12:00:42 (10588): Guest Log: "pilotErrorDiag": "Transform not found:/bin/bash: Sim_tf.py: command not found\n"

for more information: https://lhcathome.cern.ch/lhcathome/result.php?resultid=277530075

same here, on 2 machines so far:

2020-06-13 12:00:42 (10588): Guest Log: "pilotErrorDiag": "Transform not found:/bin/bash: Sim_tf.py: command not found\n"

for more information: https://lhcathome.cern.ch/lhcathome/result.php?resultid=277530075

I now tried ATLAS on a third computer - same problem, the task failed after 12 minutes :-(

see here:
https://lhcathome.cern.ch/lhcathome/result.php?resultid=277581917

can anyone tell me what causes this problem?

Hello.

Is ATLAS running fine now?

The last time I tried ATLAS was 5 days ago and that task ran fine.

I have six Atlas tasks running on a new PC with an Intel i5 CPU. I had ordered a HP desktop with an AMD Ryzen 5 3500 CPU, which would have given me 8 cores but they sent me an Intel CPU. It's a long time from the Intel PII Deschutes i had used in the Nineties. It has a 8 GB RAM and I brought it to 12 GB putting a second 4 GB RAM on the second slot. It has a 128 GB SSD disk plus a 1 TB hard disk. Wait and see.
Tullio

---