first backup is very big and incremental take a long time

[notonlyuser]

In my server I have a LSI controller with two SSD hard disks configured as Raid1. I use Debian as my operating system. The capacity from the disk is round about 900 GB:

Code: Select all

# fdisk -l /dev/sda

Disk /dev/sda: 893.3 GiB, 959119884288 bytes, 1873281024 sectors
Units: sectors of 1 * 512 = 512 bytes
Sector size (logical/physical): 512 bytes / 512 bytes
I/O size (minimum/optimal): 512 bytes / 512 bytes
Disklabel type: dos
Disk identifier: 0x000815b0

Device     Boot   Start        End    Sectors   Size Id Type
/dev/sda1          2048    8390656    8388609     4G 82 Linux swap / Solaris
/dev/sda2       8392704    9441280    1048577   512M 83 Linux
/dev/sda3       9443328 1873278976 1863835649 888.8G 83 Linux

The hard disk is filled to 17 percent.

Code: Select all

# df -h |grep sda
/dev/sda3       875G  138G  694G  17% /
/dev/sda2       488M   34M  429M   8% /boot

The first backup took a very long time and is larger than the source:

Code: Select all

Duration:           01:30:58         Processed:          244 GB (100%)
Processing rate:    46.4 MB/s        Read:               244 GB
Bottleneck:         Source           Transferred:        139 GB (1.8x)
06:51:01        Job Vollbackup started at 2017-07-21 06:51:01 CEST
06:51:02        Preparing to backup
06:51:02        Creating volume snapshot
06:51:02        Starting full backup to Repository_1
06:51:03        Backing up BIOS bootloader on /dev/sda
07:34:05        Backed up sda 244 GB at 87.1 MB/s
08:21:52        Backing up summary.xml
08:21:57        Releasing snapshot
08:21:59        Processing finished at 2017-07-21 08:21:59 CEST

The next incremental backup also took a long time, although only a few data changed.

Code: Select all

Duration:           00:52:13         Processed:          22 GB (100%)
Processing rate:    7.2 MB/s         Read:               22 GB
Bottleneck:         Source           Transferred:        3.1 GB (7.1x)

Time            Action
14:24:03        Job Vollbackup started at 2017-07-21 14:24:03 CEST
14:24:03        Preparing to backup
14:24:10        Creating volume snapshot
14:24:10        Starting incremental backup to Repository_1
14:24:13        Backing up BIOS bootloader on /dev/sda
15:13:52        Backed up sda 22 GB at 160.8 MB/s
15:16:12        Backing up summary.xml
15:16:14        Releasing snapshot
15:16:16        Processing finished at 2017-07-21 15:16:16 CEST

What can I do to speed up the backup? On another identical server, the incremental backup takes only a few minutes.

Regards,
Ronny

[nielsengelen]

Looking at the bottleneck it might be related due to the CPU however best would be to contact support for more insight on these problems.

[PTide]

Hi,

Source bottleneck means that the agent service running on the systems waits when the disk will provide data for backup. If the problem was with CPU then the bottleneck would show "Agent". Regarding long incremental run I have two questions:

- did any of the following take place between full and incremental: sever reboot, veeamsnap module unload?
- was the system under heavy IO during incremental run?

Thank you.