I'm pleased to announce that the first Robinhood User Group will take
place next September in Paris.
It will be held the afternoon before the LAD on September 21st, 2015, in
Marriott hotel, Champs Elysees, Paris.
The purpose of this event is to share experiences about Robinhood Policy
Engine
Hello,
With the new lustre.org and wiki.lustre.org site there is an opportunity
to make this a more useful tool for community involvement with Lustre.
Here are my ideas for how to arrange the new site. I know Chris Morrone
and OpenSFS members are interested in what the community can do with t
Quick question/survey:
What is the partition size folks use for their OSTs and why?
Brian Andrus
ITACS/Research Computing
Naval Postgraduate School
Monterey, California
voice: 831-656-6238
___
lustre-discuss mailing list
lustre-discuss@lists.lustre.org
For ldiskfs OSTs, I believe the most common configuration is RAID 6 with 8
disks (6+2 parity) with whatever disk size you're using.
ZFS OSTs can easily get much, much bigger, but I don't know what sizes are
typical there.
From: lustre-discuss [lustre-discuss-boun
On 04/28/2015 11:58 AM, Scott Nolin wrote:
Hello,
With the new lustre.org and wiki.lustre.org site there is an opportunity
to make this a more useful tool for community involvement with Lustre.
Here are my ideas for how to arrange the new site. I know Chris Morrone
and OpenSFS members are inter
Our biggest is a raidz6 on 10 disks of 4TB each. 8 is a good divisor
into 1MB, and the 10 disks is a good divisor as well for our total
number of disks.
bob
On 4/28/2015 4:01 PM, Andrus, Brian Contractor wrote:
Quick question/survey:
What is the partition size folks use for their OSTs and
I am running into a problem with a MDTEST benchmark with a single directory.
It appears I hit a file limit per directory using a single directory but a
unique directory test completes successfully. While running the metadata
benchmark I ran into the following error message.
MDS Server log
Thank you for clarifying the OpenSFS vs lustre.org thing Chris, this
makes sense to me.
That makes it even easier. So my suggestions now mostly boil down to
getting links to category information in the sidebar. I'm guessing that
can either be done manually by a site admin or with a mediawiki
> On Apr 28, 2015, at 4:05 PM, Patrick Farrell wrote:
>
> For ldiskfs OSTs, I believe the most common configuration is RAID 6 with 8
> disks (6+2 parity) with whatever disk size you're using.
Actually, it is probably more like 8+2 RAID6 with 128 KB chunk size (so that a
1 MB I/O request fits
Excuse me, sorry - I meant to say 8+2, like Rick said. Don't know where my
head is at.
From: Mohr Jr, Richard Frank (Rick Mohr) [rm...@utk.edu]
Sent: Tuesday, April 28, 2015 3:27 PM
To: Patrick Farrell
Cc: Andrus, Brian Contractor; lustre-discuss@lists.lus
We try to do 8+2 where it's convenient, but when we end up with 12 disk
raid arrays in some places we do 10+2.
Andrew Wagner
Research Systems Administrator
Space Science and Engineering
University of Wisconsin
andrew.wag...@ssec.wisc.edu | 608-261-1360
On 04/28/2015 03:01 PM, Andrus, Brian Cont
We like to keep the OSTs as large as is reasonable, because each OST
uses up a fixed amount of ram on _every_ Lustre client node. At our
scale, we are talking hundreds of megabytes lost on every client when we
mount lustre. With one older version of lustre I think we lost over
2GiB of RAM per
Sure, no problem! I just added a Monitoring section to the sidebar for you.
Chris
On 04/28/2015 01:25 PM, Scott Nolin wrote:
Thank you for clarifying the OpenSFS vs lustre.org thing Chris, this
makes sense to me.
That makes it even easier. So my suggestions now mostly boil down to
getting lin
I’d like to request a Testing section where I can migrate some of the
wiki.old.lustre.org content to and begin updating it.
- Justin
On 4/28/15, 3:50 PM, "Christopher J. Morrone" wrote:
>Sure, no problem! I just added a Monitoring section to the sidebar for
>you.
>
>Chris
>
>On 04/28/2015 01:2
The sidebar is just a list of links to pages in the wiki. Anyone that
wants to have a high level topic page should just go ahead and make the
relevant wiki page, much like how Scott created the
http://wiki.lustre.org/Monitoring page.
If you make a page that is a general enough starting point
Chris, after looking at my page, I think I found a better Mediawiki-ish
way to do it. I redirected my page to the category 'Monitoring'. So now
if anyone tags a page it shows up there.
I'll work on moving some other information I have over and encourage
anyone else to do the same.
Scott
On
We normally use 8+2 in RAID6 with 128KB chunk.
Andrew, what is stripe size in case of your 10+2 config? Sometime I use
256KB to make it 2.5MB stripe because I think 1.25MB is unreasonable.
On Wed, Apr 29, 2015 at 5:35 AM, Andrew Wagner
wrote:
> We try to do 8+2 where it's convenient, but when
Hi,
I would like to know which OFED, RHEL or Mellanox, do you use for MDS or
OSS and why? Is there any advantage of Mellanox OFED over RHEL OFED? I
actually experienced some sudden reset of servers while using Mellanox OFED.
I am currently use RHEL OFED for MDS and OSS. In clients, it is based on
We currently use the IB packages provided by RHEL 6 for our clients. Our
servers appear to be what’s provided by CentOS 6, but they were also part of a
shrink-wrapped solution from a vendor.
For a variety of reasons that have nothing to do with lustre, we plan on moving
the clients back to OFED
19 matches
Mail list logo
