Is there a safe way to move the bluestore wal and db to a new device
that doesn't involve rebuilding the entire OSD?
___
ceph-users mailing list -- ceph-users@ceph.io
To unsubscribe send an email to ceph-users-le...@ceph.io
If I remember correctly, being able to configure the rocksdb level
sizes was targeted for Octopus.
I was wondering if this feature ever made it into the code as it would
be useful when you want to use a drive smaller than 300G for the
WAL/DB.
___
ceph-us
Hi all,
I am attempting to prevent bluestore rocksdb Level 3/4 spillover with
a 150GB logical volume for the db/wal.
I am thinking of setting max_bytes_for_level_base to about 1.3G
(1342177280). This should let Level 3 fill up the 150GB logical
volume. I don't expect to ever actually need L4.
An
Hi all,
Does the ceph manager prometheus module export bluestore rocksdb
compaction times per OSD? I couldn't find anything.
thx
Frank
___
ceph-users mailing list -- ceph-users@ceph.io
To unsubscribe send an email to ceph-users-le...@ceph.io
That is my experience as well. The full sync will only run after
initiating a data or metadata sync init.
On Thu, Jun 11, 2020 at 9:30 AM wrote:
>
> > What I am curious about is these 2 lines:
> > full sync: 0/64 shards
> > full sync: 0/128 shards
> >
> > Is this considered normal? If so, why ha
I have seen this when there is one OSD on the node being rebooted that
is using more space than the others. Max avail for the pool is based
on the fullest OSD as far as I know.
On Sun, Jun 14, 2020 at 4:29 PM KervyN wrote:
>
> Does someone got any ideas on this?
>
> The mgr nodes are separate, pg
I believe it is the number of WALs that should be reused and should be
equal to write_buffer_number but don't quote me.
On Sat, Jun 20, 2020 at 11:35 AM Seena Fallah wrote:
>
> Hi. I found a default rocksdb option in bluestore that I can't find in
> facebook rocksdb.
> recycle_log_file_num this c
esn't change, so the I/O for metadata might be avoided (also
depends on file system mount options). Assuming most WAL files will
have similar sizes, I/O needed for metadata will be minimal.
"
On Sat, Jun 20, 2020 at 12:43 PM Frank R wrote:
>
> I believe it is the number of WALs that
log file, random writes
>> are used from size 0. Before writes hit the end of the file, the file
>> size doesn't change, so the I/O for metadata might be avoided (also
>> depends on file system mount options). Assuming most WAL files will
>> have similar sizes, I/O n
un 20, 2020 at 12:55 PM Seena Fallah wrote:
>
> Do you have any reason for this value? :))
>
> On Sat, Jun 20, 2020 at 9:24 PM Frank R wrote:
>>
>> With ceph I have always used it to set the number of WALs to recycle,
>> ie to recycle 8 WALs I use:
>>
>> &qu
fyi, there is an interesting note on disabling the write cache here:
https://yourcmc.ru/wiki/index.php?title=Ceph_performance&mobileaction=toggle_view_desktop#Drive_cache_is_slowing_you_down
On Wed, Jun 24, 2020 at 9:45 AM BenoƮt Knecht wrote:
>
> Hi Igor,
>
> Igor Fedotov wrote:
> > for the sak
I have an all RBD pool/cluster. I am interested in tracking how much disk
space is being used by each RBD image on every OSD drive.
The OSDs are Filestore.
Does anyone know of any existing scripts that accomplish this task?
If not, what commands can be used to generate this info?
___
It seems that with Linux kernel 4.16.10 krdb clients are seen as Jewel
rather than Luminous. Can someone tell me which kernel version will be seen
as Luminous as I want to enable the Upmap Balancer.
___
ceph-users mailing list -- ceph-users@ceph.io
To uns
will 4.13 also work for cephfs?
On Mon, Aug 26, 2019 at 2:31 PM Ilya Dryomov wrote:
> On Mon, Aug 26, 2019 at 8:25 PM wrote:
> >
> > What will actually happen if an old client comes by, potential data
> damage - or just broken connections from the client?
>
> The latter (with "libceph: ... feat
Would RBD performance be hurt by having thousands of cephx users defined?
___
ceph-users mailing list -- ceph-users@ceph.io
To unsubscribe send an email to ceph-users-le...@ceph.io
I have inherited a cluster where about 30% of the osds in a pool are 7200
SAS. The other 70% are 7200 SATA.
Should I look into creating 2 pools or will this likely not be a huge deal?
___
ceph-users mailing list -- ceph-users@ceph.io
To unsubscribe send
I am looking to change an RGW multisite deployment so that the secondary
will become master. This is meant to be a permanent change.
Per:
https://docs.ceph.com/docs/mimic/radosgw/multisite/
I need to:
1. Stop RGW daemons on the current master end.
On a secondary RGW node:
2. radosgw-admin zone
I have an rgw index pool that is alerting as "large" in 2 of the 3 osds on
the PG. The primary has a large omap. The index is definitely in use by the
bucket. Any opinions on the best way to solve this?
1. Remove the 2 osds with large index from cluster and rebalance?
2. Delete 2 of the 3 and deep
Hi all,
Has anyone successfully created multiple partitions on an NVME device
using ceph-disk?
If so, which commands were used?
___
ceph-users mailing list -- ceph-users@ceph.io
To unsubscribe send an email to ceph-users-le...@ceph.io
Hi all,
I have 40 nvme drives with about 20G free space each.
Would creating a 10GB partition/lvm on each of the nvmes for an rgw index
pool be a bad idea?
RGW has about about 5 million objects
I don't think space will be an issue but I am worried about the 10G size,
is it just too small for a
Is it ok to create a new OSD using ceph-volume on a server where the other
OSDs were created with ceph-disk?
thx
Frank
___
ceph-users mailing list -- ceph-users@ceph.io
To unsubscribe send an email to ceph-users-le...@ceph.io
Hi all,
After an RGW upgrade from 12.2.7 to 12.2.12 for RGW multisite a few days
ago the "sync status" has constantly shown a few "recovering shards", ie:
-
# radosgw-admin sync status
realm 8f7fd3fd-f72d-411d-b06b-7b4b579f5f2f (prod)
zonegroup 60a2cb75-6978-46a3-b830-061c8b
Hi all,
I have an issue where all "ceph pg" command hang forever, ie:
ceph pg dump
ceph pg ls-by-pool
etc
All pgs report as active+clean.
Any pointers on correcting this?
thx
Frank
___
ceph-users mailing list -- ceph-users@ceph.io
To unsubscribe send
Please disregard this. There were actually some remapped PGs.
Thx,
Frank
On Sat, Oct 26, 2019 at 12:05 PM Frank R wrote:
> Hi all,
>
> I have an issue where all "ceph pg" command hang forever, ie:
>
> ceph pg dump
> ceph pg ls-by-pool
> etc
>
> All pgs rep
Hi all,
Apologies for all the messages to the list over the past few days.
After an upgrade from 12.2.7 to 12.2.12 (inherited cluster) for an RGW
multisite
active/active setup I am almost constantly seeing 1-10 "recovering shards"
when running "radosgw-admin sync status", ie:
--
# rados
buckets. The sync error log doesn't report anything recent.
Do you know of a way to determine what this error is?
thanks
Frank
On Mon, Oct 28, 2019 at 4:38 AM Konstantin Shalygin wrote:
> On 10/27/19 6:01 AM, Frank R wrote:
> > I hate to be a pain but I have one more question.
&
seemed to clear out the shards and things are running normally now.
radosgw-admin stale instances list
didn't find these for some reason.
thx
Frank
On Wed, Oct 30, 2019 at 2:55 AM Konstantin Shalygin wrote:
> On 10/29/19 10:56 PM, Frank R wrote:
> > oldest incremental change not ap
I have a 12.2.12 cluster with 3 mons where mgr will be active on 1.
I have noticed that the command "ceph pg dump" hangs on all mons except the
one where the mgr is running.
"ceph pg dump" also runs fine on osd nodes.
Is this expected behavior?
thx
Frank
Hi all,
When the cluster map changes say due to a failed OSD how are the clients
with established sessions notified?
thx
Frank
___
ceph-users mailing list -- ceph-users@ceph.io
To unsubscribe send an email to ceph-users-le...@ceph.io
Hi all,
When increasing the number of placement groups for a pool by a large amount
(say 2048 to 4096) is it better to go in small steps or all at once?
This is a filestore cluster.
Thanks,
Frank
___
ceph-users mailing list -- ceph-users@ceph.io
To uns
Hi,
For your response:
"You should use not more 1Gb for WAL and 30Gb for RocksDB. Numbers !
3,30,300 (Gb) for block.db is useless.
"
Do you mean the block.db size should be 3, 30 or 300GB and nothing else?
If so, thy not?
Thanks,
Frank
___
ceph-users
Hi all,
How is the following situation handled with bluestore:
1. You have a 200GB OSD (no separate DB/WAL devices)
2. The metadata grows past 30G for some reason and wants to create a 300GB
level but can't?
Where is the metadata over 30G stored?
___
c
; written to main device as well.
>
>
> Thanks,
>
> Igor
> On 12/4/2019 6:13 PM, Frank R wrote:
>
> Hi all,
>
> How is the following situation handled with bluestore:
>
> 1. You have a 200GB OSD (no separate DB/WAL devices)
> 2. The metadata grows past 30G for
Is the public_network definition in ceph.conf just used to determine which
interface or IP to use for the public network or does it need to need to
encompass the public ip addresses of all cluster nodes?
Specifically, can the public_network be defined differently for different
OSD nodes as long as
Hi all,
Does anyone know what possible issues can arise if the ceph mgr daemon is
running on a mon node that has 2 ips in the public net range (1 is a
loopback address).
As I understand the it. mgr will bind to all ips
FYI - I am not sure why the loopback is there, I am trying to find out.
thx
Hi all,
I am thinking about converting a Filestore cluster to Bluestore.
The OSD nodes have 16X4TB 7200 SATA OSDs with NVME write journals. The NVME
drives should be large enough to house ~30G DB/WAL OSDs.
I am worried that I will see a significant performance hit when the
deferred writes to the
I have about 1TB of data in the pool default.rgw.logs.
What logs are stored in this pool and can they be safely deleted?
thx
Frank
___
ceph-users mailing list -- ceph-users@ceph.io
To unsubscribe send an email to ceph-users-le...@ceph.io
I am seeing the following errors on an RGW multisite slave:
1. ERROR: failed to fetch mdlog info
2. failed to fetch local sync status: (5) Input/output error
Data seems to be replicating but not metadata. Does anyone have any
ideas on what may be wrong?
-
# radosgw-admin sync status
Hi all,
This is a self inflicted issue but I am wondering if there is a way to recover.
Setup:
- RGW Multisite with 1 realm, 1 zonegroup and 2 zones.
- Metadata and data replication enable (data bi-directional).
- Only the master side is currently used by clients.
- Ceph 12.2.12
My mistake was t
Hi all,
It looks like I have an RGW multisite setup that I need to rebuild to
get metadata syncing again (I did some stupid things to break it).
If it possible to remove the slave zone from the zonegroup and then
re-add it without destroying the rgw data pool (bucket data)?
thx
Frank
___
Hi all,
When upgrading from Luminous to Nautilus the global configmap options
for cluster_network and public_network were inadvertently set to an
incorrect value (10.192.80.0/24):
-
[root@ceph-osd134 ceph]# ceph config dump | grep network
global advanced cluster_network
Hi All,
Sorry for the repost.
How do you unset a global config setting from the centralized config
with mimic+ (specifically public_network and cluster_network)
"ceph config rm global public_network"
doesn't seem to do the trick.
These were set inadvertently during an upgrade with:
"ceph conf
Hi all,
Does using the Upmap balancer require that all OSDs be the same size
(per device class)?
thx
Frank
___
ceph-users mailing list -- ceph-users@ceph.io
To unsubscribe send an email to ceph-users-le...@ceph.io
Hi all,
I really hope this isn't seen as spam. I am looking to find a position
where I can focus on Linux storage/Ceph. If anyone is currently
looking please let me know. Linkedin profile frankritchie.
Thanks,
Frank
___
ceph-users mailing list -- ceph-u
Hi all,
Is there a way to estimate how much storage space is required for
CephFS metadata given an expected number of files in the filesystem?
thx
Frank
___
ceph-users mailing list -- ceph-users@ceph.io
To unsubscribe send an email to ceph-users-le...@c
Hi all,
I have noticed that RedHat is willing to support 2x replication with
NVME drives. Additionally, I have seen CERN presentation where they
use a 2x replication with NVME for a hyperconverged/HPC/CephFS
solution.
I would like to hear some opinions on whether this is really a good
idea for pr
Thanks for clearing that up.
On Wed, Feb 19, 2020 at 9:47 AM Dan van der Ster wrote:
>
> Hi,
>
> 2x replication was for a performance test. We use 3x in production.
>
> -- dan
>
> On Wed, Feb 19, 2020 at 3:18 PM Frank R wrote:
> >
> > Hi all,
> >
&g
47 matches
Mail list logo