Hi all,
I have recently performed a few tasks, namely purging several buckets from
our RGWs and added additional hosts into Ceph causing some data movement for
a rebalance. As this is now almost completed, I kicked off some deep scrubs
and one PG is now returning the following information:
Hi all. Quick question about osd metadata information. I have several OSDs
setup with the data dir on HDD and the db going to a partition on ssd. But
when I look at the metadata for all the OSDs, it's showing the db as "hdd".
Does this effect anything? And is there anyway to change it?
$ sudo
I've used a crush location hook script to handle this before device
classes existed.
It checked the device type on startup and assigned the crush position
based on this.
I don't have that crush map any longer anywhere but the basic version
of it looked like
this: two roots "hdd" and "ssd". The
Hi everyone,
In luminous we added the crush device classes that automagically
categorize your OSDs and hdd, ssd, etc, and allow you write CRUSH rules
that target a subset of devices. Prior to this it was necessary to make
custom edits to your CRUSH map with parallel hierarchies for each
OSD
I am looking for an API equivalent of 'radosgw-admin log list' and
'radosgw-admin log show'. Existing /usage API only reports bucket level
numbers like 'radosgw-admin usage show' does. Does anyone know if this is
possible from rest API?
Thanks.
Jin.
Le vendredi 21 septembre 2018 à 19:45 +0200, Paul Emmerich a écrit :
> The cache tiering has nothing to do with the PG of the underlying
> pool
> being incomplete.
> You are just seeing these requests as stuck because it's the only
> thing trying to write to the underlying pool.
I agree, It was
On Fri, Sep 21, 2018 at 09:03:15AM +0200, Hervé Ballans wrote:
> Hi MJ (and all),
>
> So we upgraded our Proxmox/Ceph cluster, and if we have to summarize the
> operation in a few words : overall, everything went well :)
> The most critical operation of all is the 'osd crush tunables optimal', I
The cache tiering has nothing to do with the PG of the underlying pool
being incomplete.
You are just seeing these requests as stuck because it's the only
thing trying to write to the underlying pool.
What you need to fix is the PG showing incomplete. I assume you
already tried reducing the
So I've totally disable cache-tiering and overlay. Now OSD 68 & 69 are
fine, no more blocked.
But OSD 32 is still blocked, and PG 37.9c still marked incomplete with
:
"recovery_state": [
{
"name": "Started/Primary/Peering/Incomplete",
"enter_time": "2018-09-21
According to the query output you pasted shards 1 and 2 are broken.
But, on the other hand EC profile (4+2) should make it possible to recover
from 2 shards lost simultanously...
pt., 21 wrz 2018 o 16:29 Olivier Bonvalet napisał(a):
> Well on drive, I can find thoses parts :
>
> - cs0 on OSD 29
I have ubuntu servers.
With ukuu I installed kernel 4.8.17-040817 (The last < 4.9 available
kernel) and I haven't any 0x6706be76 crc since.
Nor any inconsistence.
On 19/09/18 12:01, Alfredo Daniel Rezinovsky wrote:
Tried 4.17 with the same problem
Just downgraded to 4.8. Let's see if no
Well on drive, I can find thoses parts :
- cs0 on OSD 29 and 30
- cs1 on OSD 18 and 19
- cs2 on OSD 13
- cs3 on OSD 66
- cs4 on OSD 0
- cs5 on OSD 75
And I can read thoses files too.
And all thoses OSD are UP and IN.
Le vendredi 21 septembre 2018 à 13:10 +, Eugen Block a écrit :
> > > I
Hi
I’m using 10.2.10
Thx
On Fri, Sep 21, 2018 at 9:14 AM Mykola Golub
wrote:
> Vikas, could you tell what version do you observe this on?
>
> Because I can reproduce this only on jewel, and it has been fixed
> starting since luminous 12.2.1 [1].
>
> [1] http://tracker.ceph.com/issues/20426
>
>
Yep :
pool 38 'cache-bkp-foo' replicated size 3 min_size 2 crush_rule 26
object_hash rjenkins pg_num 128 pgp_num 128 last_change 585369 lfor
68255/68255 flags hashpspool,incomplete_clones tier_of 37 cache_mode
readproxy target_bytes 209715200 hit_set
bloom{false_positive_probability: 0.05,
Hi Hendrik,
thank you for reporting the issue. I've opened a tracker issue for that,
see https://tracker.ceph.com/issues/36109.
As workaround manually configure host and port via CLI using "ceph
dashboard set-rgw-api-host " and "ceph dashboard set-rgw-api-port
"?
Regards
Volker
Am 18.09.2018
Cluster names are deprecated, don't use them. I think they might have
been removed with ceph-deploy 2.x (?)
Paul
Am Fr., 21. Sep. 2018 um 15:13 Uhr schrieb Joshua Chen
:
>
> Hi all,
> I am using ceph-deploy 2.0.1 to create my testing cluster by this command:
>
> ceph-deploy --cluster
I also switched the cache tier to "readproxy", to avoid using this
cache. But, it's still blocked.
You could change the cache mode to "none" to disable it. Could you
paste the output of:
ceph osd pool ls detail | grep cache-bkp-foo
Zitat von Olivier Bonvalet :
In fact, one object (only
Vikas, could you tell what version do you observe this on?
Because I can reproduce this only on jewel, and it has been fixed
starting since luminous 12.2.1 [1].
[1] http://tracker.ceph.com/issues/20426
On Wed, Sep 19, 2018 at 03:48:44PM -0400, Jason Dillaman wrote:
> Thanks for reporting this
Hi all,
I am using ceph-deploy 2.0.1 to create my testing cluster by this command:
ceph-deploy --cluster pescadores new --cluster-network 100.109.240.0/24
--public-network 10.109.240.0/24 cephmon1 cephmon2 cephmon3
but the --cluster pescadores (name of the cluster) doesn't seem to work.
I tried to flush the cache with "rados -p cache-bkp-foo cache-flush-
evict-all", but it blocks on the object
"rbd_data.f66c92ae8944a.000f2596".
This is the object that's stuck in the cache tier (according to your
output in https://pastebin.com/zrwu5X0w). Can you verify if that block
Could you, please paste the output of pg 37.9c query
pt., 21 wrz 2018 o 14:39 Olivier Bonvalet napisał(a):
> In fact, one object (only one) seem to be blocked on the cache tier
> (writeback).
>
> I tried to flush the cache with "rados -p cache-bkp-foo cache-flush-
> evict-all", but it blocks on
In fact, one object (only one) seem to be blocked on the cache tier
(writeback).
I tried to flush the cache with "rados -p cache-bkp-foo cache-flush-
evict-all", but it blocks on the object
"rbd_data.f66c92ae8944a.000f2596".
So I reduced (a lot) the cache tier to 200MB, "rados -p
On Fri, Sep 21, 2018 at 6:48 AM Glen Baars wrote:
>
> Hello Ceph Users,
>
>
>
> We have been using ceph-iscsi-cli for some time now with vmware and it is
> performing ok.
>
>
>
> We would like to use the same iscsi service to store our Hyper-v VMs via
> windows clustered shared volumes. When we
Hi Glen,
Yes you need clustered SCSI-3 persistent reservations support.
This is supported in SUSE SLE kernels, you may also be interested in
PetaSAN:
http://www.petasan.org
which is based on these kernels.
Maged
On 21/09/18 12:48, Glen Baars wrote:
Hello Ceph Users,
We have been using
Ok, so it's a replica 3 pool, and OSD 68 & 69 are on the same host.
Le vendredi 21 septembre 2018 à 11:09 +, Eugen Block a écrit :
> > cache-tier on this pool have 26GB of data (for 5.7TB of data on the
> > EC
> > pool).
> > We tried to flush the cache tier, and restart OSD 68 & 69, without
>
cache-tier on this pool have 26GB of data (for 5.7TB of data on the EC
pool).
We tried to flush the cache tier, and restart OSD 68 & 69, without any
success.
I meant the replication size of the pool
ceph osd pool ls detail | grep
In the experimental state of our cluster we had a cache tier
Thanks!
I was in the process of upgrading, so "noout" was already set, probably
preventing setting "noin".
I thus just "ceph osdset noup", then "ceph osd down ", which
stopped activity on the disks (probably not enough to clean everything
in Bluestore, but I decided to trust its inner working).
I
Hi,
cache-tier on this pool have 26GB of data (for 5.7TB of data on the EC
pool).
We tried to flush the cache tier, and restart OSD 68 & 69, without any
success.
But I don't see any related data on cache-tier OSD (filestore) with :
find /var/lib/ceph/osd/ -maxdepth 3 -name '*37.9c*'
I
Hello Ceph Users,
We have been using ceph-iscsi-cli for some time now with vmware and it is
performing ok.
We would like to use the same iscsi service to store our Hyper-v VMs via
windows clustered shared volumes. When we add the volume to windows failover
manager we get a device is not ready
On Thu, Sep 20, 2018 at 7:04 PM solarflow99 wrote:
>
> oh, was that all it was... git clone https://github.com/ceph/ceph-ansible/
> I installed the notario package from EPEL,
> python2-notario-0.0.11-2.el7.noarch and thats the newest they have
Hey Ken, I thought the latest versions were
Hi,
You won't be able to stop them, but if the OSDs are still running I
would just set them as out, wait for all data to be moved from them
and then it should be safe to power off the host.
---
Alex
On Fri, Sep 21, 2018 at 11:50 AM Nicolas Huillard wrote:
>
> Hi all,
>
> One of my server
Hi Olivier,
what size does the cache tier have? You could set cache-mode to
forward and flush it, maybe restarting those OSDs (68, 69) helps, too.
Or there could be an issue with the cache tier, what do those logs say?
Regards,
Eugen
Zitat von Olivier Bonvalet :
Hello,
on a Luminous
Hi,
you'll need to set `noup` to prevent OSDs from being started
automatically. The `noin` flags prevents that the cluster sets the OSD
`in` again, after it has been set `out`.
`ceph osd set noup` before `ceph osd down `
`ceph osd set noin` before `ceph osd out `
Those global flags
Hi all,
One of my server crashed its root filesystem, ie. the currently open
shell just says "command not found" for any basic command (ls, df,
mount, dmesg, etc.)
ACPI soft power-off won't work because it needs scripts on /...
Before I reset the hardware, I'd like to cleanly stop the OSDs on
Hi Cephers,
Could someone explain me how dynamic bucket index sharding works?
I have created a test bucket with 4 million objects on ceph 12.2.8 and
it showed 80 shards (ver, master_ver, max_marker fomr 0 to 79 in bucket
stats) and leave it for a night. Next day in the morning I found this in
Hi,
Am 21.09.18 um 03:28 schrieb ST Wong (ITSC):
> Hi,
>
>>> Will the RAID 6 be mirrored to another storage in remote site for DR
>>> purpose?
>>
>> Not yet. Our goal is to have the backup ceph to which we will replicate
>> spread across three different buildings, with 3 replicas.
>
> May I
Hi Hervé!
Thanks for the detailed summary, much appreciated!
Best,
MJ
On 09/21/2018 09:03 AM, Hervé Ballans wrote:
Hi MJ (and all),
So we upgraded our Proxmox/Ceph cluster, and if we have to summarize the
operation in a few words : overall, everything went well :)
The most critical
Hi MJ (and all),
So we upgraded our Proxmox/Ceph cluster, and if we have to summarize the
operation in a few words : overall, everything went well :)
The most critical operation of all is the 'osd crush tunables optimal',
I talk about it in more detail after...
The Proxmox documentation is
38 matches
Mail list logo