Hi Bryan and Dan,
Had some similar observations, and wanted some data points from you as well if possible.
1. When you say down osd , is it down and in or down and out.
2. I see the osd accumulating map ranges, when continuous recovery is going on, due to osd flaps( that is some portion of the total of pgs in cluster are not in active+clean) state.
3.Another observation is that, although it may be one of the pools that see a osd churn, the map ranges seem to increase across all the OSd across all the pools.
4.I use the asok socket to dump the status, on osd Daemon which shows the map ranges.
5.Even I tried restarting Mon, one after another, and that seemed to reduce the space, but in a sporadic fashion.
6. Once there was some breathing room, in terms of OSd not flapping, and all PG being in active + clean state, the size trimmed down to almost optimal spaces.
7. I found some pointers at :-
8. Th behaviour I am specifying us in 12.X Luminous release.
9. Would you care to share what commands you use to dump maps usage info from space perspective.
Thanks
Romit
On Mon, 9 Dec 2019, 22:56 , <ceph-users-request@xxxxxxx> wrote:
Send ceph-users mailing list submissions to
ceph-users@xxxxxxx
To subscribe or unsubscribe via email, send a message with subject or
body 'help' to
ceph-users-request@xxxxxxx
You can reach the person managing the list at
ceph-users-owner@xxxxxxx
When replying, please edit your Subject line so it is more specific
than "Re: Contents of ceph-users digest..."
Today's Topics:
1. RGW listing millions of objects takes too much time (Arash Shams)
2. Re: RGW listing millions of objects takes too much time
(Robert LeBlanc)
3. ceph mgr daemon multiple ip addresses (Frank R)
4. Re: osdmaps not trimmed until ceph-mon's restarted (if cluster has a down osd)
(Bryan Stillwell)
----------------------------------------------------------------------
Date: Mon, 9 Dec 2019 15:46:04 +0000
From: Arash Shams <ara4sh@xxxxxxxxxxx>
Subject: RGW listing millions of objects takes too much
time
To: "ceph-users@xxxxxxx" <ceph-users@xxxxxxx>
Message-ID: <CWXP265MB128679DE9E6B2A6CA289B16492580@xxxxxxxxxxxxxxxxx
P265.PROD.OUTLOOK.COM>
Content-Type: multipart/alternative; boundary="_000_CWXP265MB128679
DE9E6B2A6CA289B16492580CWXP265MB1286GBRP_"
--_000_CWXP265MB128679DE9E6B2A6CA289B16492580CWXP265MB1286GBRP_
Content-Type: text/plain; charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable
Dear All,
I have almost 30 million objects and I want to list them and index them som=
ewhere else,
Im using boto3 with continuation Marker but it takes almost 9 hours
can I run it in multiple threads to make it faster? what solution do you su=
ggest to speedup this process,
Thanks
--_000_CWXP265MB128679DE9E6B2A6CA289B16492580CWXP265MB1286GBRP_
Content-Type: text/html; charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable
<html>
<head>
<meta http-equiv=3D"Content-Type" content=3D"text/html; charset=3Diso-8859-=
1">
<style type=3D"text/css" style=3D"display:none;"> P {margin-top:0;margin-bo=
ttom:0;} </style>
</head>
<body dir=3D"ltr">
<div style=3D"font-family: Calibri, Helvetica, sans-serif; font-size: 12pt;=
color: rgb(0, 0, 0);">
Dear All, </div>
<div style=3D"font-family: Calibri, Helvetica, sans-serif; font-size: 12pt;=
color: rgb(0, 0, 0);">
<br>
</div>
<div style=3D"font-family: Calibri, Helvetica, sans-serif; font-size: 12pt;=
color: rgb(0, 0, 0);">
I have almost 30 million objects and I want to list them and index them som=
ewhere else, </div>
<div style=3D"font-family: Calibri, Helvetica, sans-serif; font-size: 12pt;=
color: rgb(0, 0, 0);">
Im using boto3 with continuation Marker but it takes almost 9 hours</div>
<div style=3D"font-family: Calibri, Helvetica, sans-serif; font-size: 12pt;=
color: rgb(0, 0, 0);">
<br>
</div>
<div style=3D"font-family: Calibri, Helvetica, sans-serif; font-size: 12pt;=
color: rgb(0, 0, 0);">
can I run it in multiple threads to make it faster? what solution do you su=
ggest to speedup this process, </div>
<div style=3D"font-family: Calibri, Helvetica, sans-serif; font-size: 12pt;=
color: rgb(0, 0, 0);">
<br>
</div>
<div style=3D"font-family: Calibri, Helvetica, sans-serif; font-size: 12pt;=
color: rgb(0, 0, 0);">
<br>
</div>
<div style=3D"font-family: Calibri, Helvetica, sans-serif; font-size: 12pt;=
color: rgb(0, 0, 0);">
Thanks </div>
<div style=3D"font-family: Calibri, Helvetica, sans-serif; font-size: 12pt;=
color: rgb(0, 0, 0);">
<br>
</div>
</body>
</html>
--_000_CWXP265MB128679DE9E6B2A6CA289B16492580CWXP265MB1286GBRP_--
------------------------------
Date: Mon, 9 Dec 2019 08:23:55 -0800
From: Robert LeBlanc <robert@xxxxxxxxxxxxx>
Subject: Re: RGW listing millions of objects takes too
much time
To: Arash Shams <ara4sh@xxxxxxxxxxx>
Cc: "ceph-users@xxxxxxx" <ceph-users@xxxxxxx>
Message-ID:
<CAANLjFoZu8LE0eH2vB9QArAOZJf4Ofm8V=jAW-+ZPOFS-f9Oag@xxxxxxxxxxxxxx>
Content-Type: multipart/alternative;
boundary="0000000000000e664e059947d3b2"
--0000000000000e664e059947d3b2
Content-Type: text/plain; charset="UTF-8"
On Mon, Dec 9, 2019 at 7:47 AM Arash Shams <ara4sh@xxxxxxxxxxx> wrote:
> Dear All,
>
> I have almost 30 million objects and I want to list them and index them
> somewhere else,
> Im using boto3 with continuation Marker but it takes almost 9 hours
>
> can I run it in multiple threads to make it faster? what solution do you
> suggest to speedup this process,
>
>
> Thanks
>
I've thought about indexing objects elsewhere as well. One thought I had
was hooking into the HTTP flow where a PUT or DEL would update the objects
in some kind of database (async of course). We could also gather stats with
GET and POST. Initially, my thoughts were to hook into haproxy since we
already use it, but possibly RGW if that is an option. That way it would
always be up to date and not have to do big scans on the buckets (our
buckets would not perform well with this). I haven't actually gotten to the
implementation phase of this idea.
----------------
Robert LeBlanc
PGP Fingerprint 79A2 9CA4 6CC4 45DD A904 C70E E654 3BB2 FA62 B9F1
--0000000000000e664e059947d3b2
Content-Type: text/html; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable
<div dir=3D"ltr"><div dir=3D"ltr">On Mon, Dec 9, 2019 at 7:47 AM Arash Sham=
s <<a href="" href="mailto:ara4sh@xxxxxxxxxxx" rel="noreferrer noreferrer" target="_blank">ara4sh@xxxxxxxxxxx">ara4sh@xxxxxxxxxxx</a>> wrot=
e:<br></div><div class=3D"gmail_quote"><blockquote class=3D"gmail_quote" st=
yle=3D"margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padd=
ing-left:1ex">
<div dir=3D"ltr">
<div style=3D"font-family:Calibri,Helvetica,sans-serif;font-size:12pt;color=
:rgb(0,0,0)">
Dear All,=C2=A0</div>
<div style=3D"font-family:Calibri,Helvetica,sans-serif;font-size:12pt;color=
:rgb(0,0,0)">
<br>
</div>
<div style=3D"font-family:Calibri,Helvetica,sans-serif;font-size:12pt;color=
:rgb(0,0,0)">
I have almost 30 million objects and I want to list them and index them som=
ewhere else,=C2=A0</div>
<div style=3D"font-family:Calibri,Helvetica,sans-serif;font-size:12pt;color=
:rgb(0,0,0)">
Im using boto3 with continuation Marker but it takes almost 9 hours</div>
<div style=3D"font-family:Calibri,Helvetica,sans-serif;font-size:12pt;color=
:rgb(0,0,0)">
<br>
</div>
<div style=3D"font-family:Calibri,Helvetica,sans-serif;font-size:12pt;color=
:rgb(0,0,0)">
can I run it in multiple threads to make it faster? what solution do you su=
ggest to speedup this process,=C2=A0</div>
<div style=3D"font-family:Calibri,Helvetica,sans-serif;font-size:12pt;color=
:rgb(0,0,0)">
<br>
</div>
<div style=3D"font-family:Calibri,Helvetica,sans-serif;font-size:12pt;color=
:rgb(0,0,0)">
<br>
</div>
<div style=3D"font-family:Calibri,Helvetica,sans-serif;font-size:12pt;color=
:rgb(0,0,0)">
Thanks=C2=A0</div></div></blockquote><div class=3D"gmail_quote"><br></div><=
div class=3D"gmail_quote">I've thought about indexing objects elsewhere=
as well. One thought I had was hooking into the HTTP flow where a PUT or D=
EL would update the objects in some kind of database (async of course). We =
could also gather stats with GET and POST. Initially, my thoughts were to h=
ook into haproxy since we already use it, but possibly RGW if that is an op=
tion. That way it would always be up to date and not have to do big scans o=
n the buckets (our buckets would not perform well with this). I haven't=
actually gotten to the implementation phase of this idea.</div><br clear=
=3D"all"><div><div dir=3D"ltr" class=3D"gmail_signature">----------------<b=
r>Robert LeBlanc<br></div></div><div>PGP Fingerprint 79A2 9CA4 6CC4 45DD A9=
04 =C2=A0C70E E654 3BB2 FA62 B9F1=C2=A0</div></div></div>
--0000000000000e664e059947d3b2--
------------------------------
Date: Mon, 9 Dec 2019 11:54:10 -0500
From: Frank R <frankaritchie@xxxxxxxxx>
Subject: ceph mgr daemon multiple ip addresses
To: ceph-users <ceph-users@xxxxxxxx>
Message-ID:
<CAMuVLDOhax+8RwPEK_52UfzmOc8dv5=YBH7Etsm5VmiO_WFEPg@xxxxxxxxxxxxxx>
Content-Type: multipart/alternative;
boundary="0000000000004ef36d0599483fd9"
--0000000000004ef36d0599483fd9
Content-Type: text/plain; charset="UTF-8"
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
Frank
mlovell - ceph anycast
--0000000000004ef36d0599483fd9
Content-Type: text/html; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable
<div dir=3D"ltr"><div>Hi all,</div><div><br>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).<br><br></div><div>A=
s I understand the it. mgr will bind to all ips<br><br></div><div>FYI - I a=
m not sure why the loopback is there, I am trying to find out.</div><div><b=
r></div><div>thx</div><div>Frank<br><br><br><br><br>mlovell - ceph anycast<=
br></div></div>
--0000000000004ef36d0599483fd9--
------------------------------
Date: Mon, 9 Dec 2019 17:24:27 +0000
From: Bryan Stillwell <bstillwell@xxxxxxxxxxx>
Subject: Re: osdmaps not trimmed until ceph-mon's
restarted (if cluster has a down osd)
To: Dan van der Ster <dan@xxxxxxxxxxxxxx>
Cc: Joao Eduardo Luis <joao@xxxxxxx>, "dev@xxxxxxx" <dev@xxxxxxx>,
ceph-users <ceph-users@xxxxxxx>
Message-ID: <9B145B17-6665-4254-8D6A-04A9B37389C3@xxxxxxxxxxx>
Content-Type: text/plain; charset="us-ascii"
On Nov 18, 2019, at 8:12 AM, Dan van der Ster <dan@xxxxxxxxxxxxxx> wrote:
>
> On Fri, Nov 15, 2019 at 4:45 PM Joao Eduardo Luis <joao@xxxxxxx> wrote:
>>
>> On 19/11/14 11:04AM, Gregory Farnum wrote:
>>> On Thu, Nov 14, 2019 at 8:14 AM Dan van der Ster <dan@xxxxxxxxxxxxxx> wrote:
>>>>
>>>> Hi Joao,
>>>>
>>>> I might have found the reason why several of our clusters (and maybe
>>>> Bryan's too) are getting stuck not trimming osdmaps.
>>>> It seems that when an osd fails, the min_last_epoch_clean gets stuck
>>>> forever (even long after HEALTH_OK), until the ceph-mons are
>>>> restarted.
>>>>
>>>> I've updated the ticket: https://tracker.ceph.com/issues/41154
>>>
>>> Wrong ticket, I think you meant https://tracker.ceph.com/issues/37875#note-7
>>
>> I've seen this behavior a long, long time ago, but stopped being able to
>> reproduce it consistently enough to ensure the patch was working properly.
>>
>> I think I have a patch here:
>>
>> https://github.com/ceph/ceph/pull/19076/commits
>>
>> If you are feeling adventurous, and want to give it a try, let me know. I'll
>> be happy to forward port it to whatever you are running.
>
> Thanks Joao, this patch is what I had in mind.
>
> I'm trying to evaluate how adventurous this would be -- Is there any
> risk that if a huge number of osds are down all at once (but
> transiently), it would trigger the mon to trim too many maps?
> I would expect that the remaining up OSDs will have a safe, low, osd_epoch ?
>
> And anyway I guess that your proposed get_min_last_epoch_clean patch
> is equivalent to what we have today if we restart the ceph-mon leader
> while an osd is down.
Joao,
I ran into this again today and found over 100,000 osdmaps on all 1,000 OSDs (~50 TiB of disk space used just by osdmaps). There were down OSDs (pretty regular occurrence with ~1,000 OSDs) so that matches up with what Dan found. Then when I restarted all the mon nodes twice the osdmaps started cleaning up.
I believe the steps to reproduce would look like this:
1. Start with a cluster with at least 1 down osd
2. Expand the cluster (the bigger the expansion, the more osdmaps that pile up)
3. Notice that after the expansion completes and the cluster is healthy that the old osdmaps aren't cleaned up
I would be willing to test the fix on our test cluster after 14.2.5 comes out. Could you make a build based on that release?
Thanks,
Bryan
------------------------------
Subject: Digest Footer
_______________________________________________
ceph-users mailing list -- ceph-users@xxxxxxx
To unsubscribe send an email to ceph-users-leave@xxxxxxx
%(web_page_url)slistinfo%(cgiext)s/%(_internal_name)s
------------------------------
End of ceph-users Digest, Vol 83, Issue 35
******************************************
------------------------------
------------------------------ ----------------------------- This email and any files transmitted with it are confidential and intended solely for the use of the individual or entity to whom they are addressed. If you have received this email in error, please notify the system manager. This message contains confidential information and is intended only for the individual named. If you are not the named addressee, you should not disseminate, distribute or copy this email. Please notify the sender immediately by email if you have received this email by mistake and delete this email from your system. If you are not the intended recipient, you are notified that disclosing, copying, distributing or taking any action in reliance on the contents of this information is strictly prohibited.
Any views or opinions presented in this email are solely those of the author and do not necessarily represent those of the organization. Any information on shares, debentures or similar instruments, recommended product pricing, valuations and the like are for information purposes only. It is not meant to be an instruction or recommendation, as the case may be, to buy or to sell securities, products, services nor an offer to buy or sell securities, products or services unless specifically stated to be so on behalf of the Flipkart group. Employees of the Flipkart group of companies are expressly required not to make defamatory statements and not to infringe or authorise any infringement of copyright or any other legal right by email communications. Any such communication is contrary to organizational policy and outside the scope of the employment of the individual concerned. The organization will not accept any liability in respect of such communication, and the employee responsible will be personally liable for any damages or other liability arising.
Our organization accepts no liability for the content of this email, or for the consequences of any actions taken on the basis of the information provided, unless that information is subsequently confirmed in writing. If you are not the intended recipient, you are notified that disclosing, copying, distributing or taking any action in reliance on the contents of this information is strictly prohibited.
------------------------------
------------------------------ -----------------------------
_______________________________________________ ceph-users mailing list -- ceph-users@xxxxxxx To unsubscribe send an email to ceph-users-leave@xxxxxxx
