[Ocfs2-devel] Please help me in getting OCFS2 design doc.
Sum Sha
sumsha.matrixreloaded at gmail.com
Mon May 8 08:23:26 CDT 2006
Thanks again for providing this information. I thought, we create
volumes with ASM and then format those volumes with OCFS2 and then
mount them. I probably missed it ;-)
One more thing which I'd be interested to discuss is "heartbeat
mechanism" and "self-fencing" behaviour of OCFS2.
I have read that
"An active node is deemed dead if it does not update its timestamp for
O2CB_HEARTBEAT_THRESHOLD (default=7) loops"
and
"A node self-fences if it fails to update its timestamp for
((O2CB_HEARTBEAT_THRESHOLD - 1) * 2) secs. The [o2hb-xx] kernel
thread, after every timestamp write, sets a timer to panic the system
after that duration. If the next timestamp is written within that
duration, as it should, it first cancels that timer before setting up
a new one"
Here, the first case seems to be dependent on the second one, isn't
it? If a node is not able to see other nodes' timestamp within
(O2CB_HEARTBEAT_THRESHOLD * 2) time, then it assumes one of the
following things:
1. The other node could not put timestamp within
(O2CB_HEARTBEAT_THRESHOLD - 1) * 2 time and paniced itself.
OR
2. The other node is actually dead and we give extra 2 seconds to
detect that. Are we giving these extra 2 seconds to [hb-xx] kernel
thread for detecting this scenario?
Is this true?
Thanks in advance.
Sumsha.
On 5/4/06, Sunil Mushran <Sunil.Mushran at oracle.com> wrote:
> The faq has been updated with descriptions of the various threads.
>
> ASM and OCFS2 have nothing in common. Volumes can either
> be OCFS2 or part of an ASM disk group but not both.
>
> Sum Sha wrote:
> > Thanks a bunch for the reply, which could give me some inside view
> > without going through the code. It'd have been great, if you had
> > written a doc mentioning all these details and rolled it out :)
> >
> > There are some ocfs2_wq, ocfs2_vote and ocfs2_cmt threads created,
> > when I mount a shared volume. Any details about them?
> >
> > Moreover, as you said that OCFS2 has shorter timings than that of CRS,
> > so do we solely depend on this fact (shorter timings) to prevent CRS
> > from intruding into racing during crash and split-brain situations. I
> > was just trying to understand how it works. I think CRS misscount is
> > 60 for Linux. Do we believe that by this time, OCFS2 would have taken
> > all the decisions and completed the recovery before CRS (or rather
> > CSSD) starts its ejection mechanism? Without ASM, how do we ensure
> > that underlying Volume Manager's recovery has been completed before
> > OCFS2 starts its recovery. And how does it happen when we use ASM?
> >
> > Sorry, if I am going out of the context, but just wanted to get some
> > information how they co-ordinate.
> >
> > Thanks.
> > Sumsha.
> > On 5/2/06, *Sunil Mushran* <Sunil.Mushran at oracle.com
> > <mailto:Sunil.Mushran at oracle.com>> wrote:
> >
> > The modules don't map one on one. The idea is to keep the
> > number of modules at a minimum while retaining some sense
> > modularity.
> >
> > For e.g., ocfs2_nodemanager.ko encompasses nm, heartbeat and net/tcp.
> >
> > The various kernel/worker threads may change over time.
> > For e.g., we added dlm_wq in the last drop. Again, the idea is to
> > keep the threads at a minimum while still functioning. :)
> >
> > o2cb (ocfs2 cluster stack) and crs have nothing in common. No, they
> > don't talk to each other. o2cb is almost entirely implemented in
> > kernel
> > space while crs in userspace.
> >
> > ocfs2 default timings are much shorter than crs. The idea is to
> > recover
> > dead nodes before crs times out.
> >
> > Sum Sha wrote:
> > > Thanks Sunil for this information. I tried going through the docs
> > > mentioned there, they are useful, but there are details which I
> > wanted
> > > for my learning purpose which are not in those docs. It would be
> > > great if anyone can share any good doc mentioning details about
> > > Clustering Services which OCFS2 uses. I know that clustering
> > services
> > > which OCFS2 uses has the following components:
> > >
> > > 1. Node manager
> > > 2. Heartbeat module
> > > 3. TCP Protocol module
> > > 4. DLM
> > >
> > > But once I install and configure OCFS2, I can see the following
> > > modules getting loaded:
> > >
> > > 1. configfs (with filesystem mounted)
> > > 2. ocfs2_nodemanger
> > > 3. ocfs2_dlm
> > > 4. ocfs2_dlmfs (with filesystem mounted)
> > >
> > > and the following threads created:
> > > 1. ocfs2_wq (perhaps it's the kernel worker thread)
> > > 2. ocfs2_vote (for voting requests. Is this the listener thread?)
> > > 3. ocfs2_cmt (perhaps this is the node monitoring thread?)
> > >
> > > Can't think of OCFS2 in 10g environment where we have another
> > required
> > > and reliable clustering service CRS (with 10g) running and managing
> > > Oracle's resources. Do CRS and OCFS's clustering services
> > co-ordinate
> > > in some manner or do they take actions independently on their own
> > > during crash and other conditions?
> > >
> > > Any help would be highly appreciated!!
> > >
> > > Thanks.
> > > Sumsha.
> > >
> > > On 4/28/06, *ocfs2-devel-request at oss.oracle.com
> > <mailto:ocfs2-devel-request at oss.oracle.com>
> > > <mailto: ocfs2-devel-request at oss.oracle.com
> > <mailto:ocfs2-devel-request at oss.oracle.com>>* <
> > > ocfs2-devel-request at oss.oracle.com
> > <mailto:ocfs2-devel-request at oss.oracle.com>
> > > <mailto:ocfs2-devel-request at oss.oracle.com
> > <mailto:ocfs2-devel-request at oss.oracle.com>>> wrote:
> > >
> > > Send Ocfs2-devel mailing list submissions to
> > > ocfs2-devel at oss.oracle.com
> > <mailto:ocfs2-devel at oss.oracle.com>
> > <mailto:ocfs2-devel at oss.oracle.com
> > <mailto:ocfs2-devel at oss.oracle.com>>
> > >
> > > To subscribe or unsubscribe via the World Wide Web, visit
> > > http://oss.oracle.com/mailman/listinfo/ocfs2-devel
> > > or, via email, send a message with subject or body 'help' to
> > > ocfs2-devel-request at oss.oracle.com
> > <mailto:ocfs2-devel-request at oss.oracle.com>
> > > <mailto:ocfs2-devel-request at oss.oracle.com
> > <mailto:ocfs2-devel-request at oss.oracle.com>>
> > >
> > > You can reach the person managing the list at
> > > ocfs2-devel-owner at oss.oracle.com
> > <mailto:ocfs2-devel-owner at oss.oracle.com>
> > > <mailto:ocfs2-devel-owner at oss.oracle.com
> > <mailto:ocfs2-devel-owner at oss.oracle.com>>
> > >
> > > When replying, please edit your Subject line so it is more
> > specific
> > > than "Re: Contents of Ocfs2-devel digest..."
> > >
> > >
> > > Today's Topics:
> > >
> > > 1. Re: Please help me in getting OCFS2 design doc. (Sunil
> > Mushran)
> > >
> > >
> > >
> > ----------------------------------------------------------------------
> > >
> > >
> > > Message: 1
> > > Date: Thu, 27 Apr 2006 10:03:23 -0700
> > > From: Sunil Mushran < Sunil.Mushran at oracle.com
> > <mailto:Sunil.Mushran at oracle.com>
> > > <mailto:Sunil.Mushran at oracle.com
> > <mailto:Sunil.Mushran at oracle.com>>>
> > > Subject: Re: [Ocfs2-devel] Please help me in getting OCFS2
> > design
> > > doc.
> > > To: Sum Sha <sumsha.matrixreloaded at gmail.com
> > <mailto:sumsha.matrixreloaded at gmail.com>
> > > <mailto:sumsha.matrixreloaded at gmail.com
> > <mailto:sumsha.matrixreloaded at gmail.com>>>
> > > Cc: ocfs2-devel at oss.oracle.com
> > <mailto:ocfs2-devel at oss.oracle.com>
> > <mailto:ocfs2-devel at oss.oracle.com
> > <mailto:ocfs2-devel at oss.oracle.com>>
> > > Message-ID: < 4450F95B.8000704 at oracle.com
> > <mailto:4450F95B.8000704 at oracle.com>
> > > <mailto:4450F95B.8000704 at oracle.com
> > <mailto:4450F95B.8000704 at oracle.com>>>
> > > Content-Type: text/plain; charset=ISO-8859-1; format=flowed
> > >
> > > Look into docs in ocfs2 projects page. While there
> > > is no design doc, there is a presentation which lists
> > > out the changes.
> > >
> > > http://oss.oracle.com/projects/ocfs2/documentation/
> > >
> > > Sum Sha wrote:
> > > > Hi all,
> > > > Could someone point me to design doc for OCFS2? I could
> > get OCFS
> > > V-1
> > > > design doc, but could not find any design doc for OCFS2. I
> > read
> > > > somewhere that OCFS2 has got lots of design changes as
> > comapared to
> > > > OCFS1.
> > > >
> > > > Thanks in advance.
> > > >
> > > > _______________________________________________
> > > > Ocfs2-devel mailing list
> > > > Ocfs2-devel at oss.oracle.com
> > <mailto:Ocfs2-devel at oss.oracle.com>
> > <mailto:Ocfs2-devel at oss.oracle.com
> > <mailto:Ocfs2-devel at oss.oracle.com>>
> > > > http://oss.oracle.com/mailman/listinfo/ocfs2-devel
> > > <http://oss.oracle.com/mailman/listinfo/ocfs2-devel
> > <http://oss.oracle.com/mailman/listinfo/ocfs2-devel>>
> > > >
> > >
> > >
> > >
> > > ------------------------------
> > >
> > > _______________________________________________
> > > Ocfs2-devel mailing list
> > > Ocfs2-devel at oss.oracle.com
> > <mailto:Ocfs2-devel at oss.oracle.com>
> > <mailto:Ocfs2-devel at oss.oracle.com
> > <mailto:Ocfs2-devel at oss.oracle.com>>
> > > http://oss.oracle.com/mailman/listinfo/ocfs2-devel
> > <http://oss.oracle.com/mailman/listinfo/ocfs2-devel>
> > >
> > >
> > > End of Ocfs2-devel Digest, Vol 25, Issue 10
> > > *******************************************
> > >
> > >
> > >
> >
> ------------------------------------------------------------------------
> >
> > >
> > > _______________________________________________
> > > Ocfs2-devel mailing list
> > > Ocfs2-devel at oss.oracle.com <mailto:Ocfs2-devel at oss.oracle.com>
> > > http://oss.oracle.com/mailman/listinfo/ocfs2-devel
> > >
> >
> >
>
More information about the Ocfs2-devel
mailing list