Hi Yuri,
On Fri, Nov 10, 2023 at 4:55 AM Yuri Weinstein wrote:
>
> I've updated all approvals and merged PRs in the tracker and it looks
> like we are ready for gibba, LRC upgrades pending approval/update from
> Venky.
The smoke test failure is caused by missing (kclient) patches in
Ubuntu
Hi, Experts,
we have a CephFS cluster running with 16.2.*, and enable multi active mds,
found somehow mds complain some info as below:
mds.*.bal find_exports balancer runs too long
and we already set below config
mds_bal_interval = 30
mds_bal_sample_interval = 12
and then we can
Thank you everyone for the feedback!
It's always good to hear if something gives value or not to the UI and to
users before we go ahead and start doing it.
And btw, if people are wondering whether we are short on features, the
short answer is no. Along with Multi-Cluster Management & monitoring
On 11/10/23 00:18, Frank Schilder wrote:
Hi Xiubo,
I will try to answer questions from all your 3 e-mails here together with some
new information we have.
New: The problem occurs in newer python versions when using the shutil.copy function. There is also
a function shutil.copy2 for which
On 11/9/23 23:41, Frank Schilder wrote:
Hi Xiubo,
great! I'm not sure if we observed this particular issue, but we did have the
oldest_client_tid updates not advancing message in a context that might re
related.
If this fix is not too large, it would be really great if it could be included
I've updated all approvals and merged PRs in the tracker and it looks
like we are ready for gibba, LRC upgrades pending approval/update from
Venky.
On Thu, Nov 9, 2023 at 1:31 PM Radoslaw Zarzynski wrote:
>
> rados approved!
>
> Details are here:
>
Hi Xiubo,
I will try to answer questions from all your 3 e-mails here together with some
new information we have.
New: The problem occurs in newer python versions when using the shutil.copy
function. There is also a function shutil.copy2 for which the problem does not
show up. Copy2 behaves a
On Wed, Nov 8, 2023 at 6:33 AM Travis Nielsen wrote:
>
> Yuri, we need to add this issue as a blocker for 18.2.1. We discovered this
> issue after the release of 17.2.7, and don't want to hit the same blocker in
> 18.2.1 where some types of OSDs are failing to be created in new clusters, or
>
Hi Xiubo,
great! I'm not sure if we observed this particular issue, but we did have the
oldest_client_tid updates not advancing message in a context that might re
related.
If this fix is not too large, it would be really great if it could be included
in the last Pacific point release.
Best
I meant this one: https://tracker.ceph.com/issues/55395
Ah, alright, almost forgot about that one.
Is there an "unmanaged: true" statement in this output?
ceph orch ls osd --export
No, it only contains the managed services that I configured.
Just out of curiosity, is there a
On Wed, Nov 8, 2023 at 11:10 AM Yuri Weinstein wrote:
>
> We merged 3 PRs and rebuilt "reef-release" (Build 2)
>
> Seeking approvals/reviews for:
>
> smoke - Laura, Radek 2 jobs failed in "objectstore/bluestore" tests
> (see Build 2)
> rados - Neha, Radek, Travis, Ernesto, Adam King
> rgw - Casey
___
ceph-users mailing list -- ceph-users@ceph.io
To unsubscribe send an email to ceph-users-le...@ceph.io
I meant this one: https://tracker.ceph.com/issues/55395
Is there an "unmanaged: true" statement in this output?
ceph orch ls osd --export
Just out of curiosity, is there a "service_name" in your unit.meta for
that OSD?
grep service_name /var/lib/ceph/{fsid}/osd.{id}/unit.meta
Zitat von
Hi Eugen,
I stopped one OSD (which was deployed by ceph orch before) and this is
what the MGR log says:
2023-11-09T13:35:36.941+ 7f067f1f0700 0 [cephadm DEBUG
cephadm.services.osd] osd id 96 daemon already exists
Before and after that are JSON dumps of the LVM properties of all OSDs.
It's the '#' character, everything after (including '#' itself) is cut
off. I tried with single and double quotes which also failed. But as I
already said, use a simple password and then change it within grafana.
That way you also don't have the actual password lying around in clear
text
IMHO we don't need yet another place to look for information, especially one
that some operators never see. ymmv.
>
>> Hello,
>>
>> We wanted to get some feedback on one of the features that we are planning
>> to bring in for upcoming releases.
>>
>> On the Ceph GUI, we thought it could be
I just tried it on a 17.2.6 test cluster, although I don't have a
stack trace the complicated password doesn't seem to be applied (don't
know why yet). But since it's an "initial" password you can choose
something simple like "admin", and during the first login you are
asked to change it
I believe they are working on it or want to work on it to revert from a
stretched cluster, because of the reason you mention: if the other datacenter
is totally burned down, you maybe want for the time being switch to one
datacenter setup.
Best regards,
Sake
> Op 09-11-2023 11:18 CET schreef
I tried everything at this point, even waited a hour, still no luck. Got it 1
time accidentally working, but with a placeholder for a password. Tried with
correct password, nothing and trying again with the placeholder didn't work
anymore.
So I thought to switch the manager, maybe something
Hi,
No, I don't think it's not very useful at best, and bad at worst.
In the IT organizations I've worked so far, any systems that actually store
data, were in the highest security zone, where no connection incoming/or
outgoing to the internet was allowed. Our systems couldn't even resolve
I was going through the hardware recommendations for a customer and
wanted to cite the memory section from the current docs [1]:
Setting the osd_memory_target below 2GB is not recommended. eph may
fail to keep the memory consumption under 2GB and extremely slow
performance is likely.
Do you see a high disk utilization? Are those OSDs hdd-only or at
least have their db on SSDs? I'd say the HDDs are the bottleneck.
There was a recent thread [1] where Zakhar explained nicely how many
IOPS you can expect from a hdd-only cluster. Maybe that helps.
[1]
Hi,
Currently, I'm testing Ceph v17.2.7 with NVMe. When mapping an rbd image to
physical compute host, "fio bs=4k iodepth=128 randwrite" give 150k IOPS. I have
a VM that located within the compute host, and the fio give ~40k IOPS and with
50 %iowait. I know there is a bottleneck, I'm not sure
Usually, removing the grafana service should be enough. I also have
this directory (custom_config_files/grafana.) but it's
empty. Can you confirm that after running 'ceph orch rm grafana' the
service is actually gone ('ceph orch ls grafana')? The directory
underneath
Hi,
I'd like to ask for confirmation how I understand the docs on stretch
mode [1]. It requires exact size 4 for the rule? Other sizes are not
supported/won't work, for example size 6? Are there clusters out there
which use this stretch mode?
Once stretch mode is enabled, it's not possible
Hi Yuri,
On Wed, Nov 8, 2023 at 4:10 PM Venky Shankar wrote:
>
> Hi Yuri,
>
> On Wed, Nov 8, 2023 at 2:32 AM Yuri Weinstein wrote:
> >
> > 3 PRs above mentioned were merged and I am returning some tests:
> > https://pulpito.ceph.com/?sha1=55e3239498650453ff76a9b06a37f1a6f488c8fd
> >
> > Still
Using podman version 4.4.1 on RHEL 8.8, Ceph 17.2.7
I used 'podman system prune -a -f' and 'podman volume prune -f' to cleanup
files, but this leaves a lot of files over in
/var/lib/containers/storage/overlay and a empty folder
/var/lib/ceph//custom_config_files/grafana..
Found those files
What doesn't work exactly? For me it did...
Zitat von Sake Ceph :
To bad, that doesn't work :(
Op 09-11-2023 09:07 CET schreef Sake Ceph :
Hi,
Well to get promtail working with Loki, you need to setup a
password in Grafana.
But promtail wasn't working with the 17.2.6 release, the URL was
To bad, that doesn't work :(
> Op 09-11-2023 09:07 CET schreef Sake Ceph :
>
>
> Hi,
>
> Well to get promtail working with Loki, you need to setup a password in
> Grafana.
> But promtail wasn't working with the 17.2.6 release, the URL was set to
> containers.local. So I stopped using it,
(I wrote it freehand, test before applying)
If your goal is to have a replication of 3 on a row and to be able to
switch to the secondary row, then you need 2 roles and you change the crush
rule on the pool side :
rule primary_location {
(...)
step take primary class ssd
step chooseleaf
On 11/9/23 07:35, Nizamudeen A wrote:
> On the Ceph GUI, we thought it could be interesting to show information
> regarding the community events, ceph release information
like others have already said, it's not the right place to put that
information for lots of reasons.
one more to add: putting
Hi Janek,
I don't really have a solution, but I tend to disagree that 'ceph
cephadm osd activate' looks for OSDs to create. It's specifically
stated in the docs that it's for existing OSDs to activate, and it did
work in my test environment. I also commented the tracker issue you
Hi Peter,
> On Nov 8, 2023, at 20:32, Peter wrote:
>
> Anyone experienced this can advise?
You can try:
* check for current cache status
smartctl -x /dev/sda | grep "Write cache"
* turn off write cache
smartctl -s wcache-sct,off,p /dev/sda
* check again
smartctl -x /dev/sda | grep "Write
Hi,
I don't think increasing the PGs has an impact on the OSD's memory, at
least I'm not aware of such reports and haven't seen it myself. But
your cluster could get in trouble as it already is, only 24 GB for 16
OSDs is too low. It can work (and apparently does) when everything is
calm,
Hi,
Well to get promtail working with Loki, you need to setup a password in
Grafana.
But promtail wasn't working with the 17.2.6 release, the URL was set to
containers.local. So I stopped using it, but forgot to click on save in KeePass
:(
I didn't configure anything special in Grafana, the
35 matches
Mail list logo
