Then you want separate partitions for each OSD journal. if you have 4 HDD
OSDs using this as they're journal, you should have 4x 5GB partitions on
the SSD.
On Mon, Jun 12, 2017 at 12:07 PM Deepak Naidu <dna...@nvidia.com> wrote:
> Thanks for the note, yes I know them all. It will be shared among multiple
> 3-4 HDD OSD Disks.
>
> --
> Deepak
>
> On Jun 12, 2017, at 7:07 AM, David Turner <drakonst...@gmail.com> wrote:
>
> Why do you want a 70GB journal? You linked to the documentation, so I'm
> assuming that you followed the formula stated to figure out how big your
> journal should be... "osd journal size = {2 * (expected throughput *
> filestore max sync interval)}". I've never heard of a cluster that
> requires such a large journal size. The default is there because it works
> for 99.999% of situations. I actually can't think of a use case that would
> require a larger journal than 10GB, especially on an SSD. The vast
> majority of the time the space on the SSD is practically empty. It doesn't
> fill up like a cache or anything. It's just a place that writes happen
> quickly and then quickly flushes it to the disk.
>
> Using 100% of your SSD size is also a bad idea based on how SSD's recover
> from unwritable sectors... they mark them as dead and move the data to an
> unused sector. The manufacturer overprovisions the drive in the factory,
> but you can help out by not using 100% of your available size. If you have
> a 70GB SSD and only use 5-10GB, then you will drastically increase the life
> of the SSD as a journal.
>
> If you really want to get a 70GB journal partition, then stop the osd,
> flush the journal, set up the journal partition manually, and make sure
> that /var/lib/ceph/osd/ceph-##/journal is pointing to the proper journal
> before starting it back up.
>
> Unless you REALLY NEED a 70GB journal partition... don't do it.
>
> On Mon, Jun 12, 2017 at 1:07 AM Deepak Naidu <dna...@nvidia.com> wrote:
>
>> Hello folks,
>>
>>
>>
>> I am trying to use an entire ssd partition for journal disk ie example
>> /dev/sdf1 partition(70GB). But when I look up the osd config using below
>> command I see ceph-deploy sets journal_size as 5GB. More confusing, I see
>> the OSD logs showing the correct size in blocks in the
>> /var/log/ceph/ceph-osd.x.log
>>
>> So my question is, whether ceph is using the entire disk partition or
>> just 5GB(default value of ceph deploy) for my OSD journal ?
>>
>>
>>
>> I know I can set per OSD or global OSD value for journal size in
>> ceph.conf . I am using Jewel 10.2.7
>>
>>
>>
>> ceph --admin-daemon /var/run/ceph/ceph-osd.3.asok config get
>> osd_journal_size
>>
>> {
>>
>> "osd_journal_size": "5120"
>>
>> }
>>
>>
>>
>> I tried the below, but the get osd_journal_size shows as 0, which is what
>> its set, so still confused more.
>>
>>
>>
>> http://docs.ceph.com/docs/master/rados/configuration/osd-config-ref/
>>
>>
>>
>>
>>
>> Any info is appreciated.
>>
>>
>>
>>
>>
>> PS: I search to find similar issue, but no response on that thread.
>>
>>
>>
>> --
>>
>> Deepak
>>
>>
>> ------------------------------
>> This email message is for the sole use of the intended recipient(s) and
>> may contain confidential information. Any unauthorized review, use,
>> disclosure or distribution is prohibited. If you are not the intended
>> recipient, please contact the sender by reply email and destroy all copies
>> of the original message.
>> ------------------------------
>> _______________________________________________
>> ceph-users mailing list
>> ceph-users@lists.ceph.com
>> http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com
>>
>
_______________________________________________
ceph-users mailing list
ceph-users@lists.ceph.com
http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com
