Hi,
You can make use of the upmap so you do not need to rebalance the entire
crush map every time you change the weight.
https://docs.ceph.com/en/latest/rados/operations/upmap/
Hope it helps,
Ansgar
Kristof Coucke schrieb am Mi., 21. Okt. 2020,
13:29:
> Hi,
>
> I have a cluster with 182 OS
There have been threads on exactly this. Might depend a bit on your ceph
version. We are running mimic and have no issues doing:
- set noout, norebalance, nobackfill
- add all OSDs (with weight 1)
- wait for peering to complete
- unset all flags and let the rebalance loose
Starting with nautilus
Hi Ansgar, Frank, all,
Thanks for the feedback in the first place.
In the meantime, I've added all the disks and the cluster is rebalancing
itself... Which will take ages as you've mentioned. Last week after this
conversation it was around 50% (little bit more), today it's around 44,5%.
Every day
Hi Frank,
We're having a lot of small objects in the cluster... RocksDb has issues
with the compaction causing high disk load... That's why we are performing
manual compaction...
See https://github.com/ceph/ceph/pull/37496
Br,
Kristof
Op ma 26 okt. 2020 om 12:14 schreef Frank Schilder :
> Hi
Okay, so far I figured out that the value in the Ceph dashboard is gathered
from a Metric from Prometheus (*ceph_osd_numpg*). Is there anyone here that
knows how this is populated?
Op ma 26 okt. 2020 om 12:52 schreef Kristof Coucke :
> Hi Frank,
>
> We're having a lot of small objects in the clu
Hi Kristof,
I missed that: why do you need to do manual compaction?
Best regards,
=
Frank Schilder
AIT Risø Campus
Bygning 109, rum S14
From: Kristof Coucke
Sent: 26 October 2020 11:33:52
To: Frank Schilder; a.jazdzew...@googlemail.com
Cc