The server structure consists of the A server that runs the backup, NetApp’s NAS server (B server) with the original path, and the C server in the Linux environment where the backup is stored.

It is mounted as a NAS on server A, and operates Rsync from server A to server C as a daemon (873 port).

Below are the logs that have already been backed up once and then incremental backups again. There are no logs left because nothing has been changed, only the building file list and the result log remain.

2024/05/02 20:34:15 [4144] building file list
2024/05/03 07:12:42 [4144] Number of files: 1585273
2024/05/03 07:12:42 [4144] Number of files transferred: 0
2024/05/03 07:12:42 [4144] Total file size: 1116744934745 bytes
2024/05/03 07:12:42 [4144] Total transferred file size: 0 bytes
2024/05/03 07:12:42 [4144] Literal data: 0 bytes
2024/05/03 07:12:42 [4144] Matched data: 0 bytes
2024/05/03 07:12:42 [4144] File list size: 40193084
2024/05/03 07:12:42 [4144] File list generation time: 0.086 seconds
2024/05/03 07:12:42 [4144] File list transfer time: 0.000 seconds
2024/05/03 07:12:42 [4144] Total bytes sent: 40254255
2024/05/03 07:12:42 [4144] Total bytes received: 61153
2024/05/03 07:12:42 [4144] sent 40254255 bytes  received 61153 bytes  1052.42 bytes/sec
2024/05/03 07:12:42 [4144] total size is 1116744934745  speedup is 27700.20
2024/05/03 07:12:42 [4144] sent 40254255 bytes  received 61154 bytes  total size 1116744934745

The building file list starts at 20:34 and rsync ends at 07:12 the next day. It took about 11 hours, which I think is too abnormal.

Below are the Rsync command log

[2024-05-02 20:34:15] cmd : rsync -a --chmod=go-rwx --no-o --no-g --stats --exclude-from 'c:plusdriveconfcbo_confMAIN1_DATACENTER_00008_dataexclude_pattern.txt' --exclude=%%share --log-file=C:/plusdrive/log/cbo/16/backup16_20240502_200000_total.log --delete --ignore-errors "/cygdrive/y/plusdriveorgcowork30008data/" "[email protected]::zeus_20.20.100.30_MAIN1_DATACENTER_00008_data" >> C:/plusdrive/log/cbo/16/backup16_20240502_200000_result.log

I’m not sure what the problem is.

First, we checked the network between each server and DISK I/O reading NAS data from server A.

I checked through iperf3, the network status between each server is very good, and the speed at which server A reads data from server B was measured through CrystalDiskMark. Based on SEQ1MQ1T1, read was measured as 250MB/s, and write was measured as 350MB/s.

I don’t think there’s a network and disk speed issue, what’s the issue?

I want to know why rsync is slow, and I want to know what to do to get faster.