Hi Cephers,
What are pros & cons of pg upmap? Is it safe to upmap every pg on the
cluster? Is this mechanism using more ram,cpu, etc?
It looks like using it will allow you to have perfect distribution of
data, but it will be good to known of some disadvantages. Has anybody
seen any?
--
Best
Hi again,
after two weeks I've got another inconsistent PG in same cluster. OSD's are
different from first PG, object can not be GET as well:
# rados list-inconsistent-obj 26.821 --format=json-pretty
{
"epoch": 178472,
"inconsistents": [
{
"object": {
Hi all,
i'm trying to copy or export a rados gateway data pool but i get an error:
rados cppool cnp-cer69.rgw.buckets.data cnp-cer69.rgw.buckets.data.save
WARNING: pool copy does not preserve user_version, which some apps may
rely on.
Hi all: extern "C" int rbd_discard(rbd_image_t image, uint64_t ofs,
uint64_t len)
{
librbd::ImageCtx *ictx = (librbd::ImageCtx *)image;
tracepoint(librbd, discard_enter, ictx, ictx->name.c_str(),
ictx->snap_name.c_str(), ictx->read_only, ofs, len);
if (len > std::numeric_limits::max()) {
Hi all,
I have a bucket with a vary big num_objects in rgw.none:
{
"bucket": "dyna",
"zonegroup": "84d584b4-3e95-49f8-8285-4a704f8252e3",
"placement_rule": "default-placement",
"explicit_placement": {
"data_pool": "default.rgw.buckets.data",
"data_extra_pool":
thanks gregory.
If I give a file name in 'ceph osd map' command still am getting 2 OSD
numbers , even this file has more objects. Why is it like this? can you
please explain
and one more doubt is
When a client write the object into primary OSD,
1. will that write be over then the primary OSD
Eugen: I've tried similar approaches in the past and it seems like it
won't work like that. I have to zap the entire journal disk. Also I plan
to use the configuration tunable for making the bluestore partition (wal,
db) larger than the default
On Mon, Aug 6, 2018 at 2:30 PM, Eugen Block
Hi,
How then can one upgrade journals to BlueStore when there is more than one
journal on the same disk?
if you're using one SSD for multiple OSDs the disk probably has
several partitions. So you could just zap one partition at a time and
replace the OSD. Or am I misunderstanding the
On Mon, Aug 6, 2018 at 5:36 PM Zhou Choury wrote:
>
> The mds of my cluster can't boot, crash all the time.
> The log is attached.
>
please set debug_mds=20 and try starting mds again.
> ___
> ceph-users mailing list
> ceph-users@lists.ceph.com
>
On Mon, Aug 6, 2018 at 9:10 AM Will Zhao wrote:
>
> Hi all: extern "C" int rbd_discard(rbd_image_t image, uint64_t ofs,
> uint64_t len)
> {
> librbd::ImageCtx *ictx = (librbd::ImageCtx *)image;
> tracepoint(librbd, discard_enter, ictx, ictx->name.c_str(),
> ictx->snap_name.c_str(),
It can start successfully now, I will send you detail log if it crash again.
-Original Message-
From: Yan, Zheng
Sent: Monday, August 6, 2018 9:00 PM
To: 周 威
Cc: ceph-users
Subject: Re: [ceph-users] ceph-mds can't start with assert failed
On Mon, Aug 6, 2018 at 5:36 PM Zhou Choury
On Mon, Aug 6, 2018 at 3:24 AM Dai Xiang wrote:
>
> On Thu, Aug 02, 2018 at 01:04:46PM +0200, Ilya Dryomov wrote:
> > On Thu, Aug 2, 2018 at 12:49 PM wrote:
> > >
> > > I create a rbd named dx-app with 500G, and map as rbd0.
> > >
> > > But i find the size is different with different cmd:
> > >
Hi all: extern "C" int rbd_discard(rbd_image_t image, uint64_t ofs, uint64_t
len)
{
librbd::ImageCtx *ictx = (librbd::ImageCtx *)image;
tracepoint(librbd, discard_enter, ictx, ictx->name.c_str(),
ictx->snap_name.c_str(), ictx->read_only, ofs, len);
if (len > std::numeric_limits::max()) {
Hello,
Any news? Still can't open new issues.
Thanks,
Alex
On Sun, Aug 5, 2018 at 1:50 PM Виталий Филиппов wrote:
>
> Thanks for the reply! Ok I understand :-)
>
> But the page still shows 403 by now...
>
> 5 августа 2018 г. 6:42:33 GMT+03:00, Gregory Farnum
> пишет:
>>
>> On Sun, Aug 5,
Hi @all,
I'm very new to Ceph and trying to learn how to setup a testing
environment in which we could explore the possibilities of Ceph.
For a complete beginner the documentation (URL:
http://docs.ceph.com/docs/master/) leaves some questions open. So I hope
to find someone on this list who
Hi Everyone,
Which is the best way to replace a failing (SMART Health Status:
HARDWARE IMPENDING FAILURE) OSD hard disk?
Normally I will:
1. set the OSD as out
2. wait for rebalancing
3. stop the OSD on the osd-server (unmount if needed)
4. purge the OSD from CEPH
5. physically replace the
Hi, our procedure is usually(assured that the cluster was ok the
failure, with 2 replicas as crush rule)
1.Stop the OSD process(to keep it from coming up and down and putting
load on the cluster)
2. Wait for the "Reweight" to come to 0(happens after 5 min i think -
can be set manually but i
Hi,
i'm running a cluster on Luminous(12.2.5), Ubuntu 16.04 - configuration
is 3 nodes, 6 drives each(though i have encountered this on a different
cluster, similar hardware, only the drives were HDD instead of SSD -
same usage). I have recently seen a bug(?) where one of the OSDs
suddenly
Waiting for rebalancing is considered the safest way, since it ensures
you retain your normal full number of replicas at all times. If you take
the disk out before rebalancing is complete, you will be causing some
PGs to lose a replica. That is a risk to your data redundancy, but it
might be an
Hello List,
We're planning to add a couple new OSD nodes to one of our clusters but
we've reached the point where we need to increase PG's before doing so. Our
ratio is currently 52pg's per OSD.
Based on the PG calc we need to make the following increases:
compute - 1024 => 4096
images 512 =>
On Mon, Aug 6, 2018 at 3:08 PM Jörg Kastning <
joerg.kastn...@uni-bielefeld.de> wrote:
> Hi @all,
>
Hi!
>
> I'm very new to Ceph and trying to learn how to setup a testing
> environment in which we could explore the possibilities of Ceph.
>
> For a complete beginner the documentation (URL:
>
On Fri, 2018-08-03 at 13:53 -0400, David Turner wrote:
> I came across you mentioning bucket check --fix before, but I totally
> forgot that I should be passing --bucket=mybucket with the command to
> actually do anything. I'm running this now and it seems to actually
> be doing something. My
Hi,
what is "min_size" on that pool? How many osd nodes you have in cluster and
do you use any custom crushmap?
On Wed, Aug 1, 2018 at 1:57 PM, shrey chauhan
wrote:
> Hi,
>
> I am trying to understand what happens when an OSD fails.
>
> Few days back I wanted to check what happens when an OSD
Hi cephers,
Is there a way to see what a crush map change does to the PG mappings
(i.e. what placement groups end up on what OSDs) without actually
setting the crush map (and have the map take effect)? I'm looking for
some way I could test hypothetical crush map changes without any effect
Thanks, I ll try to check if i can reproduce it. It's really sporadic and
occurs every 20-30 runs , I might check if it always occurs on the same
server , maybe an HW issue.
On Mon, 6 Aug 2018 at 06:12, Gregory Farnum wrote:
> This isn't very complete as it just indicates that something went
The mds of my cluster can't boot, crash all the time.
The log is attached.
ceph-mds.tj-ceph-mds2.log
Description: ceph-mds.tj-ceph-mds2.log
___
ceph-users mailing list
ceph-users@lists.ceph.com
http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com
Hi Sam,
I’m not 100% sure, but I think filestore will be removed in future ceph
releases, so you’ll have to upgrade at some point (Also, there are other
benefits you’ll get once you upgrade to bluestore).
Check different SSD caching systems included in Linux, that could be a good
solution for
Hi Andras,
From what I can tell you can run crushtool with --test
http://docs.ceph.com/docs/master/man/8/crushtool/
On Thu, Jul 26, 2018 at 1:55 AM Alex Gorbachev wrote:
>
> On Wed, Jul 25, 2018 at 7:07 PM, Alex Gorbachev
> wrote:
> > On Wed, Jul 25, 2018 at 6:07 PM, Alex Gorbachev
> > wrote:
> >> On Wed, Jul 25, 2018 at 5:51 PM, Jason Dillaman
> >> wrote:
> >>>
> >>>
> >>> On Wed, Jul 25, 2018 at 5:41
On Mon, Aug 6, 2018 at 8:13 PM Ilya Dryomov wrote:
>
> On Thu, Jul 26, 2018 at 1:55 AM Alex Gorbachev
> wrote:
> >
> > On Wed, Jul 25, 2018 at 7:07 PM, Alex Gorbachev
> > wrote:
> > > On Wed, Jul 25, 2018 at 6:07 PM, Alex Gorbachev
> > > wrote:
> > >> On Wed, Jul 25, 2018 at 5:51 PM, Jason
This has been my modus operandi when replacing drives.
Only having ~50 OSD’s for each drive type/pool, rebalancing can be a lengthy
process, and in the case of SSD’s, shuffling data adds unnecessary write wear
to the disks.
When migrating from filestore to bluestore, I would actually forklift
Hi , I got another failure and this time was able to investigate a bit.
1. If i delete the OSD and recreate it with the exact same setup, the OSD
boot up successfully
2., however, diffing the log between the failed run and the successful one
I noticed something odd:
I would have thought that with the write endurance on modern SSDs,
additional write wear from the occasional rebalance would honestly be
negligible? If you're hitting them hard enough that you're actually
worried about your write endurance, a rebalance or two is peanuts
compared to your normal
Hi,
We had the same issue after an update from luminous to mimic on ubuntu
18.04.
After updating to the latest version of the template
(https://github.com/ceph/ceph/blob/master/src/pybind/mgr/zabbix/zabbix_template.xml)
the issue was gone.
Before updating the template, zabbix_sender was
These SSD’s are definitely up to the task, 3-5 DWPD over 5 years, however I
mostly use an abundance of caution and try to minimize unnecessary data
movement so as not to exacerbate things.
I definitely could, I just er on the side of conservative wear.
Reed
> On Aug 6, 2018, at 11:19 AM,
Is anyone using nfs-ganesha in a rgw multi user / tenant environment?
I recently upgraded to nfs-ganesha 2.6 / luminous 12.2.7
___
ceph-users mailing list
ceph-users@lists.ceph.com
http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com
According to the instructions to upgrade a journal to BlueStore (
http://docs.ceph.com/docs/master/rados/operations/bluestore-migration/),
the OSD that uses the journal is destroyed and recreated.
I am using SSD journals, and want to use them with BlueStore. Reusing the
SSD requires zapping
Hi,
There’s something I would like to understand regarding advanced erasure coding
and the way objects take up place. Let’s say that I have 10 nodes of 4 OSDs
and an erasure coded pool set with K=6, M=2 and a crush failure domain of host.
I can technically fill up this ceph cluster until one
As long as your fault domain is host (or even rack) you're good, just take out
the entire host and recreate all osd's on it.
- Original Message -
From: "Robert Stanford"
To: "ceph-users"
Sent: Monday, August 6, 2018 8:39:07 PM
Subject: [ceph-users] Upgrading journals to BlueStore: a
Hi,
the missing "ln -snf ..." is probably related to missing LV tags. When
we had to migrate OSD journals to another SSD because of a failed SSD
we noticed the same difference to new (healthy) OSDs. Compare the tags
of your Logical Volumes to their actual UUIDs and all the other
40 matches
Mail list logo