On 30/10/2023 13.42, fedora@xxxxxxxxxxxxxx wrote:
F38 Linux e7.eyal.emu.id.au 6.5.8-200.fc38.x86_64 #1 SMP PREEMPT_DYNAMIC Fri Oct 20 15:53:48 UTC 2023 x86_64 GNU/Linux I am looking at issues with my system which is ATM degraded (6/7 raid6 devices). I am waiting for a replacement disk from seagate RMA. Unrelated, I am trying to copy a directory from an external USB3.0 HDD(ext4 on /sata) to a local RAID fs (ext4 on /data1). I see rsync in top as: PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND 734489 root 20 0 122048 29392 2284 R 98.3 0.1 28:37.88 rsync So high %CPU. This is pretty much constant. I am also running iostat of both the source(sdh) and target(md127) devices, and it looks like this: 12:47:06 Device tps kB_read/s kB_wrtn/s kB_dscd/s kB_read kB_wrtn kB_dscd ___ 13:12:06 sdh 0.38 1.93 0.00 0.00 116 0 0 13:13:06 sdh 0.30 1.87 0.00 0.00 112 0 0 ___ 13:12:06 md127 4.12 0.00 16.47 0.00 0 988 0 13:13:06 md127 1.90 0.00 7.60 0.00 0 456 0 Note the very low kB_read/s. hdparm tests the external as 121.29 MB/s and the array as 748.18 MB/sec (tested now, during the rsync). This also is at this slowness for the last hour, and earlier tests had it running like this for many hours. Here is an idea of what it is doing: $ sudo ps ax -o pid,wchan=WIDE-WCHAN-COLUMN -o times -o command | grep rsync 734486 do_sys_poll 0 sudo rsync -aHSK --stats --progress --checksum-choice=none --no-compress -W /sata/backups/tapes /data1/no-backup/old-backups/ 734487 do_select 16 rsync -aHSK --stats --progress --checksum-choice=none --no-compress -W /sata/backups/tapes /data1/no-backup/old-backups/ 734488 do_select 14 rsync -aHSK --stats --progress --checksum-choice=none --no-compress -W /sata/backups/tapes /data1/no-backup/old-backups/ 734489 - 2093 rsync -aHSK --stats --progress --checksum-choice=none --no-compress -W /sata/backups/tapes /data1/no-backup/old-backups/ It is clear that two threads are mostly idle and one is very busy. Why does rsync use so much CPU? Am I using the wrong options? TIA
An update - some progress. I now killed the copying and rerun it to another (plain SATA) disk on this same machine. It is going as fast as expected. This means the issue is with the array, or an interaction between the array and rsync. However, mythtv has no problem recording to the same array, at times doing multiple concurrent programs. Should I blame rsync (or the way I use it)? Next, I will do a plain 'cp -a' between the original source (sdh on /sata) and target(md127 on /data1) to see how it goes. -- Eyal at Home (fedora@xxxxxxxxxxxxxx) _______________________________________________ users mailing list -- users@xxxxxxxxxxxxxxxxxxxxxxx To unsubscribe send an email to users-leave@xxxxxxxxxxxxxxxxxxxxxxx Fedora Code of Conduct: https://docs.fedoraproject.org/en-US/project/code-of-conduct/ List Guidelines: https://fedoraproject.org/wiki/Mailing_list_guidelines List Archives: https://lists.fedoraproject.org/archives/list/users@xxxxxxxxxxxxxxxxxxxxxxx Do not reply to spam, report it: https://pagure.io/fedora-infrastructure/new_issue
