ceph Balancer not working
pradeep suresh
32 Posts
January 13, 2024, 1:31 pmQuote from pradeep suresh on January 13, 2024, 1:31 pmHi Team,
We have a small 3 node cluster with 3 OSD, out of which 2 are 8 TB and 1 is 4 TB.
I have set the ceph balancer to crush compact. Despite that, the utilization is high on the 4 TB Disk, obviously. Even setting low weight, the rebalancing is not kicking in.
How do i make crush compact balancer to work and have balanced distribution of PG/Objects.
Thanks in advance.
root@uat-ps-node1:~# ceph balancer status
{
"active": true,
"last_optimize_duration": "0:00:00.010524",
"last_optimize_started": "Sat Jan 13 07:54:53 2024",
"mode": "crush-compat",
"optimize_result": "Some osds belong to multiple subtrees: {0: ['default~hdd', 'default'], 1: ['default~hdd', 'default'], 2: ['default~hdd', 'default'], 3: ['default~hdd', 'default'], 4: ['default~hdd', 'default'], 5: ['default~hdd', 'default'], 6: ['default~ssd', 'default'], 7: ['default~ssd', 'default'], 8: ['default~tier2ssd', 'default'], 9: ['default~tier2ssd', 'default'], 10: ['default~tier2ssd', 'default'], 11: ['default~ssd', 'default']}",
"plans": []
}
ID CLASS WEIGHT REWEIGHT SIZE RAW USE DATA OMAP META AVAIL %USE VAR PGS STATUS
6 ssd 2.00000 1.00000 3.5 TiB 2.6 TiB 2.6 TiB 61 KiB 5.4 GiB 909 GiB 74.59 3.20 65 up
11 ssd 6.98630 1.00000 7.0 TiB 2.6 TiB 2.6 TiB 59 KiB 5.2 GiB 4.4 TiB 37.29 1.60 65 up
7 ssd 6.98630 1.00000 7.0 TiB 2.6 TiB 2.6 TiB 59 KiB 5.4 GiB 4.4 TiB 37.29 1.60 65 up
TOTAL 36 TiB 8.4 TiB 8.4 TiB 367 KiB 25 GiB 28 TiB 23.30
MIN/MAX VAR: 0.00/3.20 STDDEV: 23.77
Hi Team,
We have a small 3 node cluster with 3 OSD, out of which 2 are 8 TB and 1 is 4 TB.
I have set the ceph balancer to crush compact. Despite that, the utilization is high on the 4 TB Disk, obviously. Even setting low weight, the rebalancing is not kicking in.
How do i make crush compact balancer to work and have balanced distribution of PG/Objects.
Thanks in advance.
root@uat-ps-node1:~# ceph balancer status
{
"active": true,
"last_optimize_duration": "0:00:00.010524",
"last_optimize_started": "Sat Jan 13 07:54:53 2024",
"mode": "crush-compat",
"optimize_result": "Some osds belong to multiple subtrees: {0: ['default~hdd', 'default'], 1: ['default~hdd', 'default'], 2: ['default~hdd', 'default'], 3: ['default~hdd', 'default'], 4: ['default~hdd', 'default'], 5: ['default~hdd', 'default'], 6: ['default~ssd', 'default'], 7: ['default~ssd', 'default'], 8: ['default~tier2ssd', 'default'], 9: ['default~tier2ssd', 'default'], 10: ['default~tier2ssd', 'default'], 11: ['default~ssd', 'default']}",
"plans": []
}
ID CLASS WEIGHT REWEIGHT SIZE RAW USE DATA OMAP META AVAIL %USE VAR PGS STATUS
6 ssd 2.00000 1.00000 3.5 TiB 2.6 TiB 2.6 TiB 61 KiB 5.4 GiB 909 GiB 74.59 3.20 65 up
11 ssd 6.98630 1.00000 7.0 TiB 2.6 TiB 2.6 TiB 59 KiB 5.2 GiB 4.4 TiB 37.29 1.60 65 up
7 ssd 6.98630 1.00000 7.0 TiB 2.6 TiB 2.6 TiB 59 KiB 5.4 GiB 4.4 TiB 37.29 1.60 65 up
TOTAL 36 TiB 8.4 TiB 8.4 TiB 367 KiB 25 GiB 28 TiB 23.30
MIN/MAX VAR: 0.00/3.20 STDDEV: 23.77
ceph Balancer not working
pradeep suresh
32 Posts
Quote from pradeep suresh on January 13, 2024, 1:31 pmHi Team,
We have a small 3 node cluster with 3 OSD, out of which 2 are 8 TB and 1 is 4 TB.
I have set the ceph balancer to crush compact. Despite that, the utilization is high on the 4 TB Disk, obviously. Even setting low weight, the rebalancing is not kicking in.How do i make crush compact balancer to work and have balanced distribution of PG/Objects.
Thanks in advance.
root@uat-ps-node1:~# ceph balancer status
{
"active": true,
"last_optimize_duration": "0:00:00.010524",
"last_optimize_started": "Sat Jan 13 07:54:53 2024",
"mode": "crush-compat",
"optimize_result": "Some osds belong to multiple subtrees: {0: ['default~hdd', 'default'], 1: ['default~hdd', 'default'], 2: ['default~hdd', 'default'], 3: ['default~hdd', 'default'], 4: ['default~hdd', 'default'], 5: ['default~hdd', 'default'], 6: ['default~ssd', 'default'], 7: ['default~ssd', 'default'], 8: ['default~tier2ssd', 'default'], 9: ['default~tier2ssd', 'default'], 10: ['default~tier2ssd', 'default'], 11: ['default~ssd', 'default']}",
"plans": []
}ID CLASS WEIGHT REWEIGHT SIZE RAW USE DATA OMAP META AVAIL %USE VAR PGS STATUS
6 ssd 2.00000 1.00000 3.5 TiB 2.6 TiB 2.6 TiB 61 KiB 5.4 GiB 909 GiB 74.59 3.20 65 up
11 ssd 6.98630 1.00000 7.0 TiB 2.6 TiB 2.6 TiB 59 KiB 5.2 GiB 4.4 TiB 37.29 1.60 65 up
7 ssd 6.98630 1.00000 7.0 TiB 2.6 TiB 2.6 TiB 59 KiB 5.4 GiB 4.4 TiB 37.29 1.60 65 up
TOTAL 36 TiB 8.4 TiB 8.4 TiB 367 KiB 25 GiB 28 TiB 23.30
MIN/MAX VAR: 0.00/3.20 STDDEV: 23.77
Hi Team,
We have a small 3 node cluster with 3 OSD, out of which 2 are 8 TB and 1 is 4 TB.
I have set the ceph balancer to crush compact. Despite that, the utilization is high on the 4 TB Disk, obviously. Even setting low weight, the rebalancing is not kicking in.
How do i make crush compact balancer to work and have balanced distribution of PG/Objects.
Thanks in advance.
root@uat-ps-node1:~# ceph balancer status
{
"active": true,
"last_optimize_duration": "0:00:00.010524",
"last_optimize_started": "Sat Jan 13 07:54:53 2024",
"mode": "crush-compat",
"optimize_result": "Some osds belong to multiple subtrees: {0: ['default~hdd', 'default'], 1: ['default~hdd', 'default'], 2: ['default~hdd', 'default'], 3: ['default~hdd', 'default'], 4: ['default~hdd', 'default'], 5: ['default~hdd', 'default'], 6: ['default~ssd', 'default'], 7: ['default~ssd', 'default'], 8: ['default~tier2ssd', 'default'], 9: ['default~tier2ssd', 'default'], 10: ['default~tier2ssd', 'default'], 11: ['default~ssd', 'default']}",
"plans": []
}
ID CLASS WEIGHT REWEIGHT SIZE RAW USE DATA OMAP META AVAIL %USE VAR PGS STATUS
6 ssd 2.00000 1.00000 3.5 TiB 2.6 TiB 2.6 TiB 61 KiB 5.4 GiB 909 GiB 74.59 3.20 65 up
11 ssd 6.98630 1.00000 7.0 TiB 2.6 TiB 2.6 TiB 59 KiB 5.2 GiB 4.4 TiB 37.29 1.60 65 up
7 ssd 6.98630 1.00000 7.0 TiB 2.6 TiB 2.6 TiB 59 KiB 5.4 GiB 4.4 TiB 37.29 1.60 65 up
TOTAL 36 TiB 8.4 TiB 8.4 TiB 367 KiB 25 GiB 28 TiB 23.30
MIN/MAX VAR: 0.00/3.20 STDDEV: 23.77