Message boards : ATLAS application : Computational error - Disk Space exceeded
Message board moderation

To post messages, you must log in.

AuthorMessage
Fenshome

Send message
Joined: 1 Mar 15
Posts: 14
Credit: 833,705
RAC: 15
Message 45903 - Posted: 20 Dec 2021, 6:50:42 UTC

Hi

The projuct ran for 3 day then complained that it ran out of disk space.

What is the max. disk space required. Just upped it to 100Mbs from 20 Mbs is this enough ?

Kim
ID: 45903 · Report as offensive     Reply Quote
computezrmle
Volunteer moderator
Volunteer developer
Volunteer tester
Avatar

Send message
Joined: 15 Jun 08
Posts: 2051
Credit: 154,987,684
RAC: 135,822
Message 45904 - Posted: 20 Dec 2021, 7:57:22 UTC - in response to Message 45903.  

Guess it is this task reporting a peak disk usage of 10.20 GB:
https://lhcathome.cern.ch/lhcathome/result.php?resultid=337551581

IIRC the project configures it's task not to use more than 10 GB (per worker slot).
This includes:
- the VM image which grows during calculation
- the (RAM-)snapshot(s) (8400 MB each) written to the disk each time your BOINC client sends a suspend signal.


Hence, it's unclear what you mean here:
... max. disk space required. Just upped it to 100Mbs from 20 Mbs ...



That task suffered from lots of suspend/resume cycles which did not only overload the I/O subsystem.
They also caused essential network connections to fail:
2021-12-20 15:39:01 (3088): Guest Log: Checking CVMFS...
2021-12-20 15:39:03 (3088): Guest Log: Failed to check CVMFS, check output from cvmfs_config probe:
2021-12-20 15:39:04 (3088): Guest Log: Probing /cvmfs/atlas.cern.ch... Failed!
2021-12-20 15:39:04 (3088): Guest Log: Probing /cvmfs/atlas-condb.cern.ch... Failed!



ATLAS - especially configured as 6-core - might not be the right app for this computer.
You may first try to get a couple of weeks of experience using a 1-/2-core setup or with Theory.
ID: 45904 · Report as offensive     Reply Quote

Message boards : ATLAS application : Computational error - Disk Space exceeded


©2022 CERN