Re: Clarification on sequence of recovery and client ops after OSDs rejoin cluster (also, slow requests)

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



Hi Brad,

> On Sep 14, 2017, at 3:15 AM, Brad Hubbard <bhubbard@xxxxxxxxxx> wrote:
> 
> On Wed, Sep 13, 2017 at 8:40 PM, Florian Haas <florian@xxxxxxxxxxx> wrote:
>> Hi everyone,
>> 
>> 
>> disclaimer upfront: this was seen in the wild on Hammer, and on 0.94.7
>> no less. Reproducing this on 0.94.10 is a pending process, and we'll
>> update here with findings, but my goal with this post is really to
> 
> Just making sure it's understood Hammer is a retired release.

thanks for the reminder. We already noticed when upgrading our development cluster to 0.94.10 and experienced immediate continuous segfaults that was known and has a fix on the branch but was also met with “EOL”. 0.94.10 is completely unusable for me without that fix and I pulled that in manually.

Our upgrade policy has been _extremely_ cautious (much much more than we usually are on other things like kernels, Qemu, etc) as we’ve been bitten over the last years again and again by stability and performance issues.

We’re currently on the road to finally update to Jewel but wanted to figure out some of the kinks that we’ve been experiencing on hammer to find out whether those might already be fixed in Jewel or whether we’re on our way to find just another major stability/performance issue that (for some reason) we seem to be really good at stumbling into. ;)

So - we’re aware of the (recent?) Hammer EOL and we’ve been wanting to move to Jewel for a while already (but _are_ happy that we haven’t been bitten by some of the issue in the releases up to .6 or so) but we need to tread carefully.

If anyone cares to still assist on the current issue that would be very appreciated. We might consider upgrading to Jewel without fixing this first but unfortunately our dev/staging clusters aren’t always able to predict all performance/stability issues that we have encountered later in production.

Cheers,
Christian


--
Christian Theune · ct@xxxxxxxxxxxxxxx · +49 345 219401 0
Flying Circus Internet Operations GmbH · http://flyingcircus.io
Forsterstraße 29 · 06112 Halle (Saale) · Deutschland
HR Stendal HRB 21169 · Geschäftsführer: Christian Theune, Christian Zagrodnick

Attachment: signature.asc
Description: Message signed with OpenPGP

_______________________________________________
ceph-users mailing list
ceph-users@xxxxxxxxxxxxxx
http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com

[Index of Archives]     [Information on CEPH]     [Linux Filesystem Development]     [Ceph Development]     [Ceph Large]     [Linux USB Development]     [Video for Linux]     [Linux Audio Users]     [Yosemite News]     [Linux Kernel]     [Linux SCSI]     [xfs]


  Powered by Linux