Ubuntu-Certification

Copying large amounts of data causes file corruption and system freeze

Asked by Frank Weishaupt

Hardware:
      Dell Precision 7520 Intel i7-6820HQ, 2x8GB 2400MHz DDR4 Non-ECC SDRAM, 512GB M.2 SSD class 40, 1TB SATA hard drive

OS: Ubuntu 16.4 installed on SSD. The problem also occurs with Ubuntu 18.4 life CD.

Problem description:
     When large amounts of data are copied (about 5GB), then spurious errors occur. Sometimes the cp command crashes with a segmentation fault, sometimes it gets stuck. In all cases, the OS is unusable after about 30 seconds. The GUI freezes. Calling the console (CTL-ALT-F1) is not possible. After a reboot, Ubuntu works fine.

This happens when copying one large file or a tree of smaller files. The problem occurs reliably, when a quantity of about 5GB is copied. The problem occurs with hard disk involved as well as with SSD disk. The cp command runs into this problem. Also rsync runs into the problem.

Question information

Language:
English Edit question
Status:
Solved
For:
Ubuntu-Certification Edit question
Assignee:
No assignee Edit question
Solved by:
Frank Weishaupt
Solved:
Last query:
Last reply:
Revision history for this message
Frank Weishaupt (nn34919) said : 		#1

Questions:
1. What can I do to provide data that may help investigating the problem?
2. We have a support contract with Dell. So I can ask them, too, for investigation. How can I find out, which disk drivers are used in Ubuntu 16.4 installation or in the 18.4 live CD?

Revision history for this message
Pierre Equoy (pieq) said : 		#2

Hello Frank,

in the BIOS of your device, what is the SATA Operation mode selected? AHCI or RAID?

Revision history for this message
Frank Weishaupt (nn34919) said : 		#3

SATA is in RAID operation mode.

However, it seems to be not a software problem. The hardware memory test reported an error on bank 3, DIMM C. A memory replacement module is on the way. Most likely, this will be the cause of the problem. When the memory module is replaced, I'll inform you about the result of my tests.

Revision history for this message
Pierre Equoy (pieq) said : 		#4

Hi Frank,

good to know!

For your information, Canonical certifies devices in AHCI mode, not RAID.

Revision history for this message
Frank Weishaupt (nn34919) said : 		#5

Hardware is replaced on the laptop. The problem does not appear any more, even with RAID mode.

Thanks for your hint; I change to AHCI mode.