] -----
>
>
>
>
> Am 21.09.2010 22:47, schrieb Sunil Mushran:
>> There should have been another message possible just above the "cut
>> here"
>> saying possibly that the there were not enough credits, or something
>> about
>> a running or
There should have been another message possible just above the "cut here"
saying possibly that the there were not enough credits, or something about
a running or committing transaction.
On 09/21/2010 12:38 AM, Georg Höllrigl wrote:
> Hello,
>
> I got a crashed server when using ocfs2 on SLES10 wit
It is oopsing 6-8 secs after the mount. ;)
The stack trace does not show ocfs2/drbd. It is pointing to slub. But
you have to read that with a pinch of salt. This just could be a case
of some memory being scribbled.
On 09/17/2010 11:39 AM, Andre Nathan wrote:
> Hello
>
> I have an active-active DR
On 09/16/2010 09:06 AM, Enrique Sanchez wrote:
> Currently a large number of OCFS2 clusters and while the paperwork
> trolls have been pretty good at keeping us in pretty decent shape,
> I've been trying to create dynamic documentation of the several
> servers we maintain, right now I am hitting th
Which kernel are you using?
We have fixed this issue in mainline. We will soon have the same
fix for production kernels.
On 09/07/2010 02:06 PM, Todd Freeman wrote:
>From reading the archives I can see this issue has been hit before but
> I haven't found a resolution.
>
> I have a 50gb partit
tion. This partition is running multipath'd (each node had two
> Ethernet adapters) back into the same EL SANs.
>
> Anything else needed?
>
> -dan
>
>
>
> -Original Message-
> From: Sunil Mushran [mailto:sunil.mush...@oracle.com]
> Sent: Tuesday, September 07,
Can you describe your usage and your setup a bit?
On 09/07/2010 11:55 AM, Dan Lark wrote:
> I know this has been discussed before, but I am seeing high iowait times and
> the occasional deadlock between my two node OCFS2 cluster. I will be turning
> on "noatime" for the mount on both nodes in a
The stack points to netlink. ocfs2 does not use netlink. That it reproduced
with ocfs2 may just mean that the particular load triggers it. That's it.
On 09/03/2010 05:21 AM, Proskurin Kirill wrote:
Hello.
What we have:
2x Debian 5.0 x64 - 2.6.32-20~bpo50+1 from backports
DRBD + OCFS2 1.4.1-1
I
Can you attach some info in the bz. Like a iostat run for a minute
or more. top output, showing active processes. If the slowdown
is being observed by any one process, maybe a strace -p pid -T -ttt -o
/tmp/out
output.
On 07/28/2010 05:28 AM, Jeronimo Bezerra wrote:
>Guys,
>
> comments or advi
Have you tried mounting with data=writeback ?
On Jul 27, 2010, at 9:31 PM, wanchat padungrat wrote:
> Dear all,
>
> Not realy sure whether this is bug or not, but we found that sometimes OCFS2
> on our system do journaling a lot.
>
> (Please see screen shot below)
>
> As you can see, the IO
ocfs2 is a shared disk cfs. Meaning it expects the disk/vol to be
accessible by all nodes. Using fiber channel, iscsi, etc.
On 07/14/2010 12:03 PM, Alexander Nagel wrote:
> Hi,
>
> I'am new to ocfs2 filesystem and I have some questions about it.
>
> I installed three server according to the user g
You are attempting to remove a lock resource that is still active.
This is an app bug.
On 07/06/2010 11:51 AM, Charlie Sharkey wrote:
I'm seeing an occasional error on the nodes from a two node cluster.
Is this something I should be concerned about ?
Sles10 sp2 2.6.16.60-0.34-smp x86_64
tunefs.ocfs2 --cloned-volume
The man page has the details. Ensure you run the command only on the
cloned volume.
On Jun 30, 2010, at 3:19 AM, Devender Narula
wrote:
Hi guys
i got one Ocfs2 mount point running on RHEL 5.4 .. we are about to
configuring flashcopy to take backup of it .
What version are you running?
On 06/28/2010 03:46 PM, Garcia, Raymundo wrote:
> So.. if it is not cleaned up.. what can we do?
>
> -Original Message-
> From: Sunil Mushran [mailto:sunil.mush...@oracle.com]
> Sent: Monday, June 28, 2010 11:04 AM
> To: Patrick J. LoP
On 06/28/2010 09:37 AM, Patrick J. LoPresti wrote:
> On Mon, Jun 28, 2010 at 9:29 AM, Sunil Mushran
> wrote:
>
>> ocfs2 is a journaled file system. But it is also a clustered file system.
>> So it cannot arbitrarily delete orphaned files because they could still be
>
On 06/28/2010 07:57 AM, Patrick J. LoPresti wrote:
> On Sun, Jun 27, 2010 at 11:17 PM, Garcia, Raymundo
> wrote:
>
>> Hello… it was put under my attention that a partition we have in one of our
>> production system was displaying wrong size with df command…. 123 GB… but in
>> fact the size of
No, only a fsck will remove the two inodes from the orphan dir. And
until that's run, that message will be printed every 10 mins on some
node in the cluster.
The bug that led to this problem was fixed in 2.6.34.
http://git.kernel.org/?p=linux/kernel/git/torvalds/linux-2.6.git;a=commit;h=3939fda4b3
As far as ocfs2 is concerned, the current version of ocfs2 1.2
is ocfs2 1.2.9. You will find the packages for your kernel on
oss.oracle.com. The news section has the list of changes/bugs fixed.
asmlib also has some updates. You can review the fixes to see
whether an upgrade is warranted.
For all
For issues on sles, file a bz with Novell. Make sure you list out
the fs features that had been enabled.
On 06/15/2010 07:08 PM, Patrick J. LoPresti wrote:
> My O/S is Suse Linux Enteprise Server 11 Service Pack 1.
>
> My SCSI device is a hardware iSCSI RAID chassis. I have done a
> variety of re
- lopre...@gmail.com wrote:
> Hello. I am experimenting with OCFS2 on Suse Linux Enterprise Server
> 11 Service Pack 1.
>
> I am performing various stress tests. My current exercise involves
> writing to files using a shared-writable mmap() from two nodes.
> (Each
> node mmaps and writes
- bpkr...@gmail.com wrote:
> Patrick J. LoPresti 2010-06-13 19:14:
> > Hello. I am experimenting with OCFS2 on Suse Linux Enterprise
> Server
> > 11 Service Pack 1.
> >
> > I am performing various stress tests. My current exercise involves
> > writing to files using a shared-writable mmap
Yes, that's part of the tools. Tools are not kernel dependent. Just
install the one for rhel5 that is downloadable from oss.oracle.com.
On Jun 12, 2010, at 8:50 AM, Jeffrey Layton wrote:
Got it - thanks!
I'm looking for /etc/init.d/o2cb and I don't see it. Does that come
with ocfs2tools?
We could remove this check. If you want this in your sles kernel, the
quickest route will be via Novell. We'll need both sles and mainline
patched.
On Jun 11, 2010, at 5:54 PM, "Patrick J. LoPresti"
wrote:
> Hello. I am experimenting with OCFS2 on a brand new 10GigE iSCSI SAN.
> It looks
Other than a longer pause. No other negative side effects. Processes
that have the locks at the required level will continue to do io.
Processes that need to upconvert a lock and need the "dead" node to
respond, will have to wait for the deadthreshold to expire before
recovery can clean out
You mean the next version of 1.4 tools? Not anytime soon I'm afraid.
That patch needs to be pushed to the 1.4 branch before you can build it.
On 06/04/2010 10:42 AM, Ulf Zimmermann wrote:
> Sunil, Tao,
>
> Can you tell me an estimation when a new tools package with this patch will
> be available
On 06/04/2010 07:17 AM, Andrew Robert Nicols wrote:
> If the hang is only short, could it be that we're just missing the relevant
> busy locks by running scanlocks too late?
>
Then it is not a hang. It is just slow. A hang is more permanent
and is typically due to a bug in some component. A bu
If scanlocks is clean, means it is not a dlm issue.
Have you tried mounting with data=writeback? With drbd,
a 1G write becomes a 2G write. With ordered mode, a journal
checkpoint, which is done when relinquishing a write lock, will
wait on the data flush. That could be the cause for the slowdown.
It is not a dlm issue.
On 06/03/2010 07:38 AM, Andrew Robert Nicols wrote:
On Thu, Jun 03, 2010 at 11:12:49AM +0100, Andrew Robert Nicols wrote:
What's the best place to start looking for the cause of these hangs? I've
attached the dmesg output which includes some call traces for hung threa
t was the
"bottleneck" in the first place.
On 06/01/2010 11:00 AM, Stefan Priebe - allied internet ag wrote:
>
> No not but bonnie is still as slow as before when creating and
> deleting files. I thought this should be "fixed" when using indexed-dirs.
>
> Stef
The kernel is old. We fixed this issue in 2.6.30. We have also backported
it to the 1.4 production tree.
The problem was that the inodes being created did not have locality
leading to a directory having inodes that were spaced far apart from
each other. The one place where it really affected perfo
.32.12 kernel. Also the disk was formatted using:
> mkfs.ocfs2 --fs-feature-level=max-features -L ocfs2disk -N 10 -T mail
> -v /dev/sdb
>
> So i thought that i can now use features like unwritten, inline-data
> and indexed-dirs.
>
> Stefan
>
> Sunil Mushran schrieb:
>&g
On 06/01/2010 04:06 AM, Stefan Priebe - allied internet ag wrote:
> Feature Compat: 3 backup-super strict-journal-super
> Feature Incompat: 8016 sparse extended-slotmap inline-data
> metaecc xattr indexed-dirs refcount
> Tunefs Incomplete: 0
> Feature RO co
On 05/26/2010 01:39 PM, Daniel McDonald wrote:
>
>> ocfs2 does not reset without a log message. Do you have netconsole
>> setup? Messages logged a tick before reset can only be captured by
>> netconsole/kdump etc.
>>
> Unfortunately no. Here are the two lines in /var/log/message prior to the
When a node dies, the cluster ops pause for the node to be first
declared dead followed by recovery. Threshold governs the time
it takes to declare the node dead. The higher the value, the longer
the pause.
ocfs2 does not reset without a log message. Do you have netconsole
setup? Messages logged a
-107 means the node lost connection with the other node. The messages
below appear cut-pastes and not in sequence. So I cannot tell for sure
what happened next. What should have happened is that the node would
then go into quorum mode followed by recovery mode.
Sunil
On 05/26/2010 05:38 AM, Fran
Fragmentation has been atop our dev priority list for sometime
now. That is, both, reducing it and handling it better when it does
get fragmented.
Just last week we pushed patches for the same into the newly
created 2.6.35.
http://oss.oracle.com/pipermail/ocfs2-devel/2010-May/006511.html
As alwa
http://oss.oracle.com/projects/ocfs2-tools/news/article_8.html
We did make a related change in fsck in that release. Do you mind
creating a bugzilla for this? Do mention the arch. I can then send you
a debug version of the tool that'll tell us why it is behaving like that
on your machine.
On 05/2
for that
The server is a redhat enterprise Linux Server release 5.4 and we 're
using the following rpm
ocfs2-tools-1.4.3-1.el5
ocfs2-2.6.18-164.2.1.el5-1.4.4-1.el5
ocfs2console-1.4.3-1.el5
Le 20/05/2010 16:32, Sunil Mushran a écrit :
Versions?
On May 20, 2010, at 6:47 AM, ste
Versions?
On May 20, 2010, at 6:47 AM, stephane lomine > wrote:
Hello
Sorry if ask a question already seen before but i'm new to ocfs2
and was not able to find a proper answer on the web.
We re using ocfs2 on a two nodes system with SAN disks.
We need to extend 3 of our partitions so we a
The fs is oopsing when trying to remove a entry from the orphan dir.
It could be that that orphaned inode is corrupted. You could try
running fsck.ocfs2 - fy /dev/sdX. Better if you ping Novell support
for assistance.
On May 17, 2010, at 6:01 AM, Georg Höllrigl
wrote:
> Hi Folks,
>
> I´
On 05/10/2010 11:53 PM, Stefan Priebe - allied internet ag wrote:
> Do you mean the trunk at all?
> http://oss.oracle.com/git/?p=ocfs2-tools.git;a=summary
>
Yes.
>>> 2.) Is there a release date for 1.6?
>>>
>> We are planning to do a beta release for el6. No dates as yet.
>> Ping Nove
e cluster
> Start ocfs2
> Mount ocfs2 disks
> Start oracle cluster
>
>
> This plan will need to be done in a quiet period of the system, otherwise the
> running node will have to much pressure on it.
>
> Regards
> Morten K
>
> -Opprinnelig mel
On 05/10/2010 05:18 PM, Sunil Mushran wrote:
> Upgrade to el5 u4 kernel. 2.6.18-164.el5. ocfs2 1.2 is provided for
> all el5 kernels upto u4.
Correction. It is provided on el5 upto u3. 2.6.18-128.el5.
___
Ocfs2-users mailing list
Ocfs2
.oracle.com] *På vegne av* Sunil Mushran
*Sendt:* 7. mai 2010 19:07
*Til:* ocfs2-users@oss.oracle.com
*Emne:* Re: [Ocfs2-users] Kernel panic when deleting a file
Unsure why you have to build the packages when they are downloadable.
Please use the packages provided on oss.
On 05/07/2010 06:26 AM, Kr
On 05/10/2010 01:33 PM, Stefan Priebe - allied internet ag wrote:
> 1.) Is ocfs2-tools 1.6 considered as stable?
>
Mostly. We push changes to the tree only after it passes all the tests.
> 2.) Is there a release date for 1.6?
>
We are planning to do a beta release for el6. No dates as ye
Unsure why you have to build the packages when they are downloadable.
Please use the packages provided on oss.
On 05/07/2010 06:26 AM, Kristiansen Morten wrote:
Hi again,
This time I try attach a jpeg.
Regards
Morten K
-Opprinnelig melding-
Fra: ocfs2-users-boun...@oss.oracle.com
[ma
On 05/04/2010 02:42 AM, Werner Flamme wrote:
>
> Thank you. Took a while to get this posting out of spam quarantine :-(
> We had problems here for the last weeks :-(
>
> So, on the SLES10 where I succeeded in building the packages for ocfs2
> and ocfs2-tools from source, I entered this command and
Cannot say for sure. It could be a deadlock (bug) too. As in, I don't
want to
blame any one entity without knowing more.
If it were up to me, I'd start with the dlm. See which node holds the lock
that others are waiting on. Then see why that node is unable to downconvert
that lock. As in, if the
Werner Flamme wrote:
> this is what I do. But since the version on the RAC server is newer than
> mine, I cannot mount the filesystem (I quoted the error in a previous
> mail). That's why I try to compile the sources from Oracle. I'd rather
> deinstall the SLES version and switch over to the new
Werner Flamme wrote:
> Unfortunately, I do not run the Oracle RAC. The RAC runs with Oracle
> Unbreakable Linux. I am sorry for that decision, but I can't change it.
>
> SAP does not forbid to use Oracle's version of Linux. SAP only says that
> SAP systems are fully supported on RHEL or SLES only (
Werner Flamme wrote:
> For RHEL there are readyly built packages, why would I build my own
> packages then? The build worked fine with another SLES10.
>
> Yes, ocfs2 is included by SLES. Not SLES 11, of course - it is in the
> separately sold "High Availability" package. But the version included in
1.4 tree is only meant to be build against EL5 U2+. Not SLES
nor any other kernel tree. SLES9/10/11 already includes ocfs2.
Werner Flamme wrote:
> -BEGIN PGP SIGNED MESSAGE-
> Hash: SHA1
>
> Hi,
>
> next VM, next trouble :-(
>
> Now I work inside a VM with SLES 11. Configuring ocfs2-1.4.7
Your setup may not have persistent device naming. You can
use blkid or mounted.ocfs2 to discover the ocfs2 devices.
Marcus Alves Grando wrote:
> Hello Guys,
>
> After upgrade to 1.4.7-1 my FS does not mount anymore. Just after
> upgrade rpm, FS mount and works fine, but after reboot server, it do
gt; [mailto:ocfs2-users-boun...@oss.oracle.com] On Behalf Of Sunil Mushran
> Sent: Monday, April 19, 2010 1:08 PM
> To: ocfs2-annou...@oss.oracle.com; ocfs2-users
> Subject: [Ocfs2-users] OCFS2 1.4.7-1 and OCFS2 Tools 1.4.4-1 released
>
> All,
>
> We are pleased to announce the
res: M095a03 Owner: 18 State: 0x0
> Last Used: 0 ASTs Reserved: 0Inflight: 0Migration Pending: No
> Refs: 3Locks: 1On Lists: None
> Reference Map:
> Lock-Queue Node Level Conv Cookie Refs AST BAST
> Pending-Action
&
All,
We are pleased to announce the release of OCFS2 1.4.7-1 and OCFS2 Tools
1.4.4-1 for Oracle's and Red Hat's Enterprise Linux 5 Update 2 and higher.
Oracle's Unbreakable Linux Network users who are subscribing to the "OCFS2
1.4 packages for Enterprise Linux 5" channel can upgrade to this relea
Joel Becker wrote:
> On Thu, Apr 15, 2010 at 12:31:02PM +0200, Kristiansen Morten wrote:
>> I discovered our four node cluster running on RedHat EL5, Ocfs2 1.2.6 and
>> Oracle 10.2.0.3 have memory leak. I suspect ocfs2, but I could be wrong. I
>> suspect ocfs2 because when we run RMAN backup the
ons but I feel this
> could become an issue again.
>
> David
>
> -Original Message-
> From: ocfs2-users-boun...@oss.oracle.com
> [mailto:ocfs2-users-boun...@oss.oracle.com] On Behalf Of Sunil Mushran
> Sent: Thursday, April 15, 2010 12:14 PM
> To: li...@svrinfor
We are hoping to release it anyday now.
Have you filed a bug about your issue? I have no recollection of any
reports of such an issue. Orphan scanning has not changed in 1.4.7.
File a bz. We'll need to get more information to understand the
problem you are encountering.
Mailing List SVR wrote:
>
No. it should not be running for so long.
Attach strace to it. It could be caught in a loop.
Schildwachter, Xavier wrote:
>
> We setup four nodes connected to an iSCSI SAN with 3 ocfs2 volumes.
>
> Last Friday, one of the volumes was set read-only because of the
> following error:
>
> Apr 9 01:4
Ping Oracle Support. They will be able to answer qs on OracleVM.
You are seeing 4 paths because they might be multipathed. Check your
iscsi configuration.
Use blkid to determine which two paths are the same.
e.g.
/dev/sdf1: LABEL="label1" UUID="908a0229-88c3-4a0d-b6bc-38c43c6b1461"
TYPE="ocfs2"
beat=local)
>
> My concern was with writing status files for the other nodes to see. If
> the partition was mounted read-only, would that cause another node to think
> that the read-only node has failed?
>
> Thanks,
>
> Kevin
>
> On Mon, 05 Apr 2010 14:03:52 -0700, Sunil M
ocfs2 can handle multiple writers to the same file.
Cannot say whether it is due to the io load. All I can say
is that it is an io error.
Unsure what /data partitions are.
Sunil
ke...@utahsysadmin.com wrote:
> Sunil,
>
> Thanks for the response. Could this be triggered by both servers trying t
It is having problems doing ios to the virtual devices. -5 is EIO.
ke...@utahsysadmin.com wrote:
> I have a relatively new test environment setup that is a little different
> from your typical scenario. This is my first time using OCFS2, but I
> believe it should work the way I have it setup.
>
>
Attach it to a bugzilla. Don't cut-paste it. It is impossible to decipher.
Jason Price wrote:
> Ok. I have netconsole setup, and echo 't' > /proc/sysrq-trigger did
> not cause the box to crash.
>
> Here's the output sent over to the other host:
>
> My source host setup was:
>
> dmesg -n 8
>
> ser
If "fs_locks -B" is empty, then the processes are not waiting on a
cluster lock.
Process pegged at 100% cpu means it is actively waiting to acquire a
spinlock.
Is the other process running?
Unfortunately in EL5 there is no clean way to get the kernel stack for a
process.
"echo t >/proc/sysrq-t
Where are you downloading from?
Florin Andrei wrote:
> ##
> There was a package dependency problem. The message was:
>
> Unresolvable chain of dependencies:
> ocfs2-2.6.18-128.el5-1.4.4-1.el5 requires ocfs2-tools >= 1.4.3
>
>
> The following packages we
Yes and without the trailing nulls.
Florin Andrei wrote:
> So what would happen if we use 1.4 instead? Would the logger keep
> logging data and the user would be able to read the new lines normally
> while node A is down?
>
>
> On 03/30/2010 11:42 AM, Sunil Mushran wrote:
&g
What you are seeing is the result of writeback data journaling
in ocfs2 1.2. In ocfs2 1.4, we default to ordered data journaling.
Refer to the 1.4 user's guide for more.
Florin Andrei wrote:
> A and B are identical machines. Network has lots of redundancy. They
> both access same OCFS2 volumes ove
What's the current size of the partition? The error indicates
that the partition has not been resized.
Mattia Gandolfi wrote:
> Hi all,
>
> I'm a new OCFS2 user, I'm trying to implement a 2-nodes cluster with a
> shared fs, and I'm facing an issue while trying to resize an existing
> ocfs2 files
; mtime: 0x4a0b2372 -- Wed May 13 14:45:54 2009
> dtime: 0x0 -- Wed Dec 31 18:00:00 1969
> ctime_nsec: 0x -- 0
> atime_nsec: 0x0000 -- 0
> mtime_nsec: 0x -- 0
> Last Extblk: 0
>
No
On Mar 29, 2010, at 8:10 PM, Angelo McComis wrote:
> Does it matter that the nodes are numbered 1-6 instead of 0-5?
>
>
>
> On Mon, Mar 29, 2010 at 4:25 PM, Sunil Mushran > wrote:
>> Enable some debugging.
>>
>> #debugfs.ocfs2 -l TCP allow
>> .
102.141
> Connection to 192.168.102.141 port [tcp/cbt] succeeded!
>
> -Original Message-
> From: Sunil Mushran [mailto:sunil.mush...@oracle.com]
> Sent: Monday, March 29, 2010 5:08 PM
> To: David Murphy
> Cc: ocfs2-users@oss.oracle.com
> Subject: Re: [Ocfs2
= -107
> ocfs2: Unmounting device (253,1) on (node 0)
>
>
>
> So clearly ocfs2 the service things it can connect to the node, but nmap
> sees the connection just fine. And Web2 can see the port on web1 just fine,
> so there is no firewall blocking the connections.
If node A is a lower node number than node B, then the behavior
is correct. In a 2 node cluster, if the two nodes cannot talk to each
other, the higher node number will fence itself. Also, when a node
mounts a volume, it initiates connections to other live nodes. If any
connection fails, the mount
I checked. The core support for xattr in tools was added in 1.4.2.
We plan on pinging ubuntu to pick up tools 1.6 (when it is ready)
that has xattr support enabled by default.
Frank Lahm wrote:
> Hi,
>
> although official Oracle documentation indeed suggests that EA should
> not work if ocfs-tool
http://oss.oracle.com/projects/ocfs2/files/RedHat/RHEL5/
Devender Narula wrote:
> Hi Guys
>
> i need ocfs2 software for 2.6.18-164.15.1.el5 RHEL 64 bit.. Can
> anybody please tell me from where i can download it.
>
> thanks
>
> Devender
>
___
Oc
Please file a bugzilla. This may involve more people. And it
is easier to track that way.
David Johle wrote:
> See attachment for requested output.
>
> At 08:54 PM 3/24/2010, Sunil Mushran wrote:
>> Quite a bit of work is ongoing on this front. I'll list all that work
hmm.. o2cb_ctl makes no connections. It just reads the cluster.conf
and populates configfs. AFAIK.
David Murphy wrote:
>
> We had 6 nodes running CentOS 5.4 using 1.4.3 ocfs2-tools.
>
>
>
> I decided to rebuild one node with FC12.
>
>
>
>
>
> Which is working fine, however
>
>
>
> Nmap 1
Quite a bit of work is ongoing on this front. I'll list all that work
in another email.
Meanwhile make a bz with the stat_sysdir output. We'll need that
to determine the best way forward.
David Johle wrote:
> So in light of prior issues with lock contention and such due to
> writing apache logs
The network connect is failing. Could be because of a firewall,
or bad ip address, some switch issue.
Mount the volume on node 2. Then enable tracing and
tail messages file.
# debugfs.ocfs2 -l TCP allow
# tail -f /var/log/messages
Then from node 4, ping node 2 using netcat.
# nc -z 192.168.1.2 77
at appears to be the
> reason, why I can't remove the snapshot.
>
> Is there any (safe) way get rid of it after unmounting?
>
> Thank's
> Armin
>
>
>
>
>> -Original Message-
>> From: Sunil Mushran [mailto:sunil.mush...@oracle.com]
>>
ocfs2-tools-1.6 will have support for it. It has not been released
as yet. But you are free to build it for your use.
http://oss.oracle.com/git/?p=ocfs2-tools.git;a=summary
Redshift wrote:
> Hello,
>
> I am using the linux-image-2.6.32-trunk-amd64 Debian kernel with OCFS2
> "1.5.0" (at least, th
C: 214-587-3882
>
> -Original Message-
> From: Sunil Mushran [mailto:sunil.mush...@oracle.com]
> Sent: Thursday, March 18, 2010 1:25 PM
> To: Jaquays, Michael A.
> Cc: ocfs2-users@oss.oracle.com
> Subject: Re: [Ocfs2-users] NFS in "D" State
>
> I am assum
One option is to provide me with the o2image of the volume.
# o2image -r /dev/sda1 - | bzip2 > sda1.out.bz2
File a bugzilla and add the link to that image. (The bz cannot
handle large files.)
The other option is to file a bz and attach the stat_sysdir output.
http://oss.oracle.com/~smushran/.debu
I am assuming you are mounting the nfs mounts with the nordirplus
mount option. If not, that is known to deadlock a nfsd thread leading
to what you are seeing.
There are two possible reasons for this error. One is a dlm issue.
Other is a local deadlock like above.
To see if the dlm is the cause f
Yeah.. mounted is a bit dumb. In the next release, it will recognize
/dev/mapper devices. We still need to teach it to handle multipathing
fully.
David Johle wrote:
> I'm not sure about why mounted.ocfs2 is showing both the dm and the
> sd devices for the same volume. But this could all be very
I am queasy recommending such a setup to anyone. It is one thing to handle
a workload. The problem is about handling user/admin errors. You are
essentially
running a local volume manager that is unaware of the other node. Any
reconfig
that is not coordinated will lead to corruption. Below that yo
:1224
> Recovering node 9 from slot 7 on device (152,0)
>
> But the ocfs2 disk was unavailable anyway.
>
> Any other hint?
>
> Regards,
>
> G.
>
> On Wed, Mar 10, 2010 at 8:56 PM, Sunil Mushran
> wrote:
>> Were the first set of messages on all nodes? On that
To get this feature, one needs both a kernel >= 2.6.28 and a
functioning userspace cluster stack.
ocfs2 1.4 on (rh)el5/sles10 does not satisfy either of the two.
As of now, there are two such stacks. Pacemaker from Novell and
the "new" CMAN (new is my term... unsure how RH will be marketing
it) f
Were the first set of messages on all nodes? On that node atleast
the o2hb node down event fired. It should have fired on all nodes.
This is the dlm eviction message.
If they all fired, then look for a node to have a message that
reads "Node x is the Recovery Master for the Dead Node y".
That sho
It is waiting for the heartbeat timeout to trigger denoting a
node death. Then it initiates recovery. With the default settings,
this takes about a minute plus. This is with the o2cb cluster stack.
Note not all fs ops will hang during the detection phase. Only
those ops will hang that directly or
ocfs2 release 1.4 does not have these features. These features (expect
quotas) will
be in ocfs2 release 1.6. The 1.4 user's guide lists the features
included in that release.
Elia Pinto wrote:
> Hi
>
> I'd like to know if OCFS2 1.4 on RHEL5 with the REDHAT latest kernel
> supports disk quotas,
On Mar 2, 2010, at 2:52 AM, Brad Plant wrote:
> On Tue, 2 Mar 2010 21:51:45 +1100
> Brad Plant wrote:
>
>> On Tue, 2 Mar 2010 05:09:21 -0500
>> Enrique Sanchez wrote:
>>
> During my test (take Node0 down cold turkey) Node1 hung pretty
> badly,
> is this something expected??
Dmitry Rybin wrote:
> I have storage 2TB with 4K cluster size. Collected statistic says,
> that 4K max reading block from disc.
> But extX - variable read size up to 50K with cluster size 4K.
>
> ocfs2 have no prefetch?
>
> Centos 5.4, ocfs2-1.4.4-1
>
Is the workload the same on the two volumes
The volume goes ro when it detects an on-disk corruption.
I would imagine the node detected the problem with the group
allocator fixed by fsck. dmesg will tell you more. If it mentions
block#1709568, then fsck took care of it.
Sunil
Enrique Sanchez wrote:
> Hello folks,
>
> I have a Filesystem th
t;
> -Original Message-
> From: Sunil Mushran [mailto:sunil.mush...@oracle.com]
> Sent: Thursday, February 18, 2010 11:47 AM
> To: Jaquays, Michael A.
> Cc: ocfs2-users@oss.oracle.com
> Subject: Re: [Ocfs2-users] Kernel Panic ocfs2_inode_lock_full
>
> Yes, this is a k
That's a poor workload for a clustered file system because it has
to master/take clustered locks for each inode. Actually multiple
locks per inode. Those locks are used once and then freed. A local
fs only takes a read hit.
loebber...@eplan.de wrote:
> Hi Guy,
>
> if you have serveral OCFS2 partit
Yes, this is a known issue. Only occurs when nfs is in the equation.
This issue has been fixed in mainline quite some time ago. We are
in the process of backporting that to 1.4.
michael.a.jaqu...@verizon.com wrote:
> All,
>
> I have a 3 node cluster that is experiencing kernel panics once every fe
Brad Plant wrote:
> I noticed that there were a lot of bug fixes in 2.6.33-rc8. Just wondering if
> any of these are also applicable to 2.6.32 and if they'll be merged into the
> long term stable branch?
>
> If there are no 2.6.32 merge plans, are any of the 2.6.33 commits beneficial
> to 2.6.32
301 - 400 of 1424 matches
Mail list logo
