Hi everyone,
I have a question about the network setup. From the document, It’s recommended
to have 2 NICs per hosts as described in below picture
[Diagram]
In the picture, OSD hosts will connect to the Cluster network for replicate and
heartbeat between OSDs, therefore, we definitely need 2 NICs
Hi Frank, Patrick,
I tried to create a cephfs elsewhere and got the warning, that's what
got my attention. Nothing is broken, but I thought it could have impact
on performance.
Very good to know it's not that bad!
Thanks!!
Kenneth
On 07/05/2020 23:31, Patrick Donnelly wrote:
On Wed, Apr 2
Hello Nghia,
just use one network interface card and use frontend and backend traffic on
the same. No problem with that.
If you have a dual port card, use both ports as an LACP channel and maybe
separate it using VLANs if you want to, but not required as well.
--
Martin Verges
Managing director
If an upmap is not stored, it means that OSDMap::check_pg_upmaps is
deciding that those upmaps are invalid for some reason.
Additional debugging can help sort out why.
(Maybe you have a complex crush tree and the balancer is creating
invalid upmaps).
-- dan
On Fri, May 1, 2020 at 2:48 PM Andras P
Hi Martin,
Thanks for your response. You mean one network interface for only MON hosts or
for the whole cluster including OSD hosts? I’m confusing now because there are
some projects that only use one public network for the whole cluster. That
means the rebalancing, replicating objects and hear
Hello Nghia,
I once asked a similar question about network architecture and got the
same answer as Martin wrote from Wido den Hollander:
There is no need to have a public and cluster network with Ceph. Working
as a Ceph consultant I've deployed multi-PB Ceph clusters with a single
public netwo
Hi
Sorry for the repost, but I didn't get any respons on my first post so I
will try to rephrase it
We have several ceph clusters running nautilus (coming from mimic). On one
of the clusters I got a health warning
HEALTH_WARN
1 large omap objects
* when I check on the rados gateway
Hi,
Are there any more resources of unit tests for CRUSH algorithm other than
the test cases here: :
https://github.com/ceph/ceph/tree/master/src/test/crush
Or more unit testing of CRUSH apart from the these test cases would be an
overkill?
BR
Bobby !
Hi Marcel,
What version of nautilus is this? It sounds similar to
https://tracker.ceph.com/issues/43188, which should be fixed in
14.2.9.
On Fri, May 8, 2020 at 6:18 AM Marcel Kuiper wrote:
>
> Hi
>
> Sorry for the repost, but I didn't get any respons on my first post so I
> will try to rephras
Hi Casey,
We're running 14.2.8. I will schedule an upgrade a.s.a.p. Thanks for your
respons
Marcel
> Hi Marcel,
>
> What version of nautilus is this? It sounds similar to
> https://tracker.ceph.com/issues/43188, which should be fixed in
> 14.2.9.
>
>
> On Fri, May 8, 2020 at 6:18 AM Marcel Kuipe
Hi Dan,
You are absolutely right - it didn't occur to me to check whether the
upmaps are correct, and indeed some of them are obviously not. I do have
a more complex crush rule for our 6+3 EC pool (pool 9):
step take root-disk
step choose indep 3 type pod
step choose i
On all OSD nodes I'm using vm.min_free_kbytes = 4194304 (4GB). This was one of
the first tunings on the cluster.
Best regards,
=
Frank Schilder
AIT Risø Campus
Bygning 109, rum S14
From: Anthony D'Atri
Sent: 08 May 2020 10:17
To: Frank Sc
I’ve inherited a couple of clusters with non-default (ie, not “ceph”) internal
names, and I want to rename them for the usual reasons.
I had previously developed a full list of steps - which I no longer have access
to.
Anyone done this recently? Want to be sure I’m not missing something.
* N
Are they LVM based?
The keyring files should be just the filenames, yes.
Here's a recent list I saw which was missing the keyring step but is
reported to be complete otherwise.
- Stop RGW services
- Set the flags (noout,norecover,norebalance,nobackfill,nodown,pause)
- Stop OSD/MGR/MON services
-
14 matches
Mail list logo
