Forums

Home / Forums

You need to log in to create posts and topics. Login · Register

1 pools have many more objects per pg than average


Abhishek Kumar Singh
18 Posts
December 21, 2020, 7:10 am

Dear Team,

From today we are getting this health warnings for my ceph cluster. I can't delete any pools.

Thanks in advance

Abhishek

#1

admin
2,930 Posts
December 21, 2020, 8:59 am

it means 1 pool has many objects but similar pg count than pools with lower object count.

you could suppress ths with setting

mon_pg_warn_max_object_skew  to 0

put in in the Ceph Configuration ui under global section

the other option is to re-adjust the pg count but this will cause a lot of rebalance traffic and not recommended.

#2

Abhishek Kumar Singh
18 Posts
December 24, 2020, 10:17 am

Hi Team,

We have 3 node working cluster. we are thinking to add one more server but we are stuck at node interfaces.

Current Cluster network setting,

bond0: eth0(1G),eth1(1G)[Management]

bond1: eth5(10G),eth6(10G)[Backend]

bond2: eth2(10G),eth7(10G) [iSCSI].

But In our new server eth2 is 1G port. How can we rectify this issue.

Hope you understand my query.

Thanks in advance,

Abhishek

#3

admin
2,930 Posts
December 24, 2020, 12:13 pm

You can rename or re-order interface names from the blue screen menu on the node itself.

#4