On Tue, Aug 6, 2019 at 7:56 PM Konstantin Shalygin wrote:
> Is it possible to add a new device class like 'metadata'?
>
>
> Yes, but you don't need this. Just use your existing class with another
> crush ruleset.
>
Maybe it's the lateness of the day, but I'm not sure how to do that. Do you
have
Hi Manuel,
Yes, I already tried that option but the result it's extremely noisy and
not usable due to lack of some fields, besides that forget to parse those
logs in order to print some stats. Also, I'm not sure if this is a good
hint to rgw performance.
I think I'm going to stick with nginx and
Hi all.
I want to delay time for sync process from primary zone to secondary zone.
If some one want to delete my data i have enough time to process.
How i can do it. Config some options, install more proxy. Any solutions.
Thanks. Regards
___
ceph-user
Is it possible to add a new device class like 'metadata'?
Yes, but you don't need this. Just use your existing class with another
crush ruleset.
If I set the device class manually, will it be overwritten when the OSD
boots up?
Nope. Classes assigned automatically when OSD is created, not
On Tue, Aug 6, 2019 at 11:11 AM Paul Emmerich
wrote:
> On Tue, Aug 6, 2019 at 7:45 PM Robert LeBlanc
> wrote:
> > We have a 12.2.8 luminous cluster with all NVMe and we want to take some
> of the NVMe OSDs and allocate them strictly to metadata pools (we have a
> problem with filling up this clu
-63> 2019-08-07 00:51:52.861 7fe987e49700 1 heartbeat_map
clear_timeout 'OSD::osd_op_tp thread 0x7fe987e49700' had suicide timed
out after 150
You hit a suicide timeout, that's fatal. On line 80 the process kills
the thread based on the assumption it's hung.
src/common/HeartbeatMap.cc:
66 boo
Hi,
I think -> default.rgw.buckets.index for us it reach 2k-6K iops for a index
size of 23GB.
Regards
Manuel
-Mensaje original-
De: ceph-users En nombre de
dhils...@performair.com
Enviado el: miércoles, 7 de agosto de 2019 1:41
Para: ceph-users@lists.ceph.com
Asunto: [ceph-users] Rado
All;
Based on the PG Calculator, on the Ceph website, I have this list of pools to
pre-create for my Object Gateway:
.rgw.root
default.rgw.control
default.rgw.data.root
default.rgw.gc
default.rgw.log
default.rgw.intent-log
default.rgw.meta
default.rgw.usage
default.rgw.users.keys
default.rgw.user
All;
I have a server running CentOS 7.6 (1810), that I want to set up with CephFS
(full disclosure, I'm going to be running samba on the CephFS). I can mount
the CephFS fine when I use the option secret=, but when I switch to
secretfile=, I get an error "No such process." I installed ceph-com
> However, I'm starting to think that the problem isn't with the number
> of threads that have work to do... the problem may just be that the
> OSD & PG code has enough thread locking happening that there is no
> possible way to have more than a few things happening on a single OSD
> (or perhaps a
Hi
We got a pair of OSD located in node that crash randomly since 14.2.2
OS Version : Centos 7.6
There're a ton of lines before crash , I will unespected:
--
3045> 2019-08-07 00:39:32.013 7fe9a4996700 1 heartbeat_map is_healthy
'OSD::osd_op_tp thread 0x7fe987e49700' had timed ou
Thanks, that looks quite useful.
I did a few tests and got basically a null result. In fact, when I
put the RBDs on different pools on the same SSDs or pools on different
SSDs, performance was a few percent worse than leaving them on the
same pool. I definitely wasn't expecting this!
It looks l
> Your parallel rsync job is only getting 150 creates per second? What
> was the previous throughput?
I am actually not quite sure what the exact throughput was or is or what
I can expect. It varies so much. I am copying from a 23GB file list that
is split into 3000 chunks which are then process
On Tue, Aug 6, 2019 at 7:45 PM Robert LeBlanc wrote:
> We have a 12.2.8 luminous cluster with all NVMe and we want to take some of
> the NVMe OSDs and allocate them strictly to metadata pools (we have a problem
> with filling up this cluster and causing lingering metadata problems, and
> this w
We have a 12.2.8 luminous cluster with all NVMe and we want to take some of
the NVMe OSDs and allocate them strictly to metadata pools (we have a
problem with filling up this cluster and causing lingering metadata
problems, and this will guarantee space for metadata operations). In the
past, we hav
You may be interested in using my wallclock profiler to look at lock
contention:
https://github.com/markhpc/gdbpmp
It will greatly slow down the OSD but will show you where time is being
spent and so far the results appear to at least be relatively
informative. I used it recently when refa
On Tue, Aug 6, 2019 at 7:57 AM Janek Bevendorff
wrote:
>
>
> > 4k req/s is too fast for a create workload on one MDS. That must
> > include other operations like getattr.
>
> That is rsync going through millions of files checking which ones need
> updating. Right now there are not actually any cre
I have a few more cycles this week to dedicate to the problem of
making OSDs do more than maybe 5 simultaneous operations (as measured
by the iostat effective queue depth of the drive).
However, I'm starting to think that the problem isn't with the number
of threads that have work to do... the pro
On 08/06/2019 11:28 AM, Mike Christie wrote:
> On 08/06/2019 07:51 AM, Matthias Leopold wrote:
>>
>>
>> Am 05.08.19 um 18:31 schrieb Mike Christie:
>>> On 08/05/2019 05:58 AM, Matthias Leopold wrote:
Hi,
I'm still testing my 2 node (dedicated) iSCSI gateway with ceph 12.2.12
bef
On 08/06/2019 07:51 AM, Matthias Leopold wrote:
>
>
> Am 05.08.19 um 18:31 schrieb Mike Christie:
>> On 08/05/2019 05:58 AM, Matthias Leopold wrote:
>>> Hi,
>>>
>>> I'm still testing my 2 node (dedicated) iSCSI gateway with ceph 12.2.12
>>> before I dare to put it into production. I installed lat
Hi Felix,
You can increase debug option with debug rgw in your rgw nodes.
We got it to 10.
But at least in our case we switched again to civetweb because it don’t provide
a clear log without a lot verbose.
Regards
Manuel
De: ceph-users En nombre de Félix Barbeira
Enviad
Hi,
I'm testing radosgw with beast backend and I did not found a way to view
more information on logfile. This is an example:
2019-08-06 16:59:14.488 7fc808234700 1 == starting new request
req=0x5608245646f0 =
2019-08-06 16:59:14.496 7fc808234700 1 == req done req=0x5608245646f0
op
4k req/s is too fast for a create workload on one MDS. That must
include other operations like getattr.
That is rsync going through millions of files checking which ones need
updating. Right now there are not actually any create operations, since
I restarted the copy job.
I wouldn't expec
On Tue, Aug 6, 2019 at 12:48 AM Janek Bevendorff
wrote:
> > However, now my client processes are basically in constant I/O wait
> > state and the CephFS is slow for everybody. After I restarted the copy
> > job, I got around 4k reqs/s and then it went down to 100 reqs/s with
> > everybody waiting
Am 05.08.19 um 18:31 schrieb Mike Christie:
On 08/05/2019 05:58 AM, Matthias Leopold wrote:
Hi,
I'm still testing my 2 node (dedicated) iSCSI gateway with ceph 12.2.12
before I dare to put it into production. I installed latest tcmu-runner
release (1.5.1) and (like before) I'm seeing that bo
hi folks,
we had to move one of our clusters so we had to boot all servers, now
we found an Error on all OSD with the EC-Pool.
do we miss some opitons, will an upgrade to 13.2.6 help?
Thanks,
Ansgar
2019-08-06 12:10:16.265 7fb337b83200 -1
/build/ceph-13.2.4/src/osd/ECUtil.h: In function
'ECUti
On Mon, Aug 5, 2019 at 6:35 PM wrote:
> > Hi Team,
> > @vita...@yourcmc.ru , thank you for information and could you please
> > clarify on the below quires as well,
> >
> > 1. Average object size we use will be 256KB to 512KB , will there be
> > deferred write queue ?
>
> With the default settin
However, now my client processes are basically in constant I/O wait
state and the CephFS is slow for everybody. After I restarted the copy
job, I got around 4k reqs/s and then it went down to 100 reqs/s with
everybody waiting their turn. So yes, it does seem to help, but it
increases latency
Thanks that helps. Looks like the problem is that the MDS is not
automatically trimming its cache fast enough. Please try bumping
mds_cache_trim_threshold:
bin/ceph config set mds mds_cache_trim_threshold 512K
That did help. Somewhat. I removed the aggressive recall settings I set
before an
29 matches
Mail list logo