Concurrent login firewire disk

Hi
I'm trying to set up an Oracle RAC Test Environment concerning Mr.
Hunter's article (http://www.oracle.com/technology/pub/articles/hunter_rac_2.html).
I use Linux RedHat Fedora C2 on both nodes (2 Compaq ML370) and a
Maxtor OneTouch 200GB as external shared device. The drive is
connected through Firewire to both nodes (LaCie Firewire PCI-Card).
I also installed the precompiled RedHat Kernel provided by Oracle
(kernel-2.4.21-15.ELorafw1.i686.rpm). But unfortunately I am not able
to connect the drive from both nodes at the same time. One node works
correctly. But the other one is not able to login into the drive. If
After "modprobe ohci1394" and "modprobe sbp2 sbp2_exclusive_login=0" I
executed "dmesg" and got following messages:
Node 1 (login ok):
Query logins to SBP-2 device successful
Maximum concurrent logins supported: 3
Number of active logins: 0
Logged into SBP-2 device
Node[01:1023]: Max speed [S400] - Max payload [2048]
Node 2 (login fails):
Query logins to SBP-2 device successful
Maximum concurrent logins supported: 3
Number of active logins: 1
Error logging into SBP-2 device - login failed
If I run the "rescan-scsi-bus.sh" script, Node 1 shows the Maxtor
device under "OLD", Node 2 shows only "0 devices found".
Does anybody know this problem or had similar issues? Does anybody use
a Maxtor OneTouch as shared device successfully? Because Oracle
recommends to use this disk and because this disk has got an "Oxford
911" Chipset, I assume that the drive allows multiple logins!? Or am I
assuming wrong?
Please give me a hint. I don't know any more what I can do...
Thank you very much

I am also having the same problem. I use OneTouch II 250G drive, followed Jeff Hunter article, and have tried FireWire cards from 3 different vendors. No success.
Here is what happened:
boot node1, node1 got connection:
[root@linux1 ~]# dmesg | grep sbp2
sbp2: $Rev: 1265 $ Ben Collins <[email protected]>
ieee1394: sbp2: Driver forced to serialize I/O (serialize_io = 1)
ieee1394: sbp2: Maximum concurrent logins supported: 2
ieee1394: sbp2: Number of active logins: 1
ieee1394: sbp2: Logged into SBP-2 device
now boot node2, node2 got connection:
[root@linux2 ~]# dmesg | grep sbp2
sbp2: $Rev: 1265 $ Ben Collins <[email protected]>
ieee1394: sbp2: Driver forced to serialize I/O (serialize_io = 1)
ieee1394: sbp2: Maximum concurrent logins supported: 2
ieee1394: sbp2: Number of active logins: 1
ieee1394: sbp2: Logged into SBP-2 device
but the node1 connection is gone:
[root@linux1 ~]# dmesg | grep sbp2
sbp2: $Rev: 1265 $ Ben Collins <[email protected]>
ieee1394: sbp2: Driver forced to serialize I/O (serialize_io = 1)
ieee1394: sbp2: Maximum concurrent logins supported: 2
ieee1394: sbp2: Number of active logins: 1
ieee1394: sbp2: Logged into SBP-2 device
ieee1394: sbp2: Error reconnecting to SBP-2 device - reconnect failed
ieee1394: sbp2: Logged out of SBP-2 device
ieee1394: sbp2: Maximum concurrent logins supported: 2
ieee1394: sbp2: Number of active logins: 0
ieee1394: sbp2: Logged into SBP-2 device
I just can't get this damm firewire thing working. Any hints, tips would be highly appreciated,
Thanks

Similar Messages

Unable to make tWo concurrent logins to the firewire Hard Disk from RHEL 3.

Hello
I tried setting up a firewire(shared storage) ORACLE 10 G RAC CLUSTER using OCFS/ASM using your install guidelines
but this failed.Below is the scenario.
Hardware : compaq intel PC's with 1 GB RAM
Members in the cluster : Two
Operating system : RHEL AS 3 / FEDORA CORE -3 (both being tried one after the other)
Oracle 10G RAC version : 10.1.0.3.0
HDD : Maxtor one touch II 250 GB with combo interface Firewire/USB 2.0
N/W Adapters : Two public interconnect,Two private interconnect thru 10/100 internet Hubs/switches
Firewire card details: FireWire (IEEE 1394): Agere Systems (former Lucent Microelectronics) FW323 (rev 05)
Failure Description
I successfully loaded the drivers for the Maxtor one touch II in the two Linux (RHEL 3) boxes. But after a restart, at any point of time,
the disk is detected only by one node. It is detected in the other node only if the first node is powered OFF. On investigating we found the
following error message by running "dmesg | grep sbp2" command as root
ieee1394: sbp2: Query logins to SBP-2 device successful
ieee1394: sbp2: Maximum concurrent logins supported: 2
ieee1394: sbp2: Number of active logins: 1
ieee1394: sbp2: Error logging into SBP-2 device - login failed
Please do suggest where i was wrong and how to proceed next.
Thanks & Regards
Raghavan

Hello All,
I work as the system admin for Raghavan. Thanks to all those who took time to reply to our queries. Sorry for the lengthy mail. Let me go through the whole sequence of what we did initially.
1. Installed RHEL 3 with Agere PCI Firewire cards.
2. Installed Firewire Kernel kernel-2.4.21-27.0.2.ELorafw1 on top of it.
3. Modules were loaded properly and the Firewire hard disk was accessible from both the nodes.
4. Everything went fine during OCFS and ASM installation.
5. We were installing the CRS and that too went fine till 99% and the final step of running the root.sh in one of the nodes gave an error message as follows
"Failure at final check of Oracle CRS Stack
0"
6. So we thought of reinstalling CRS and deleted the Voting disk file directly from the shared disk. We deleted the file manually from one of the nodes using the "rm -rf" command and that too before uninstalling OCFS.
Now both the disks went out of sync. The voting disk and CRS file was still visible from the other node. So we did a reboot of both the nodes one by one. After this MYSTERIOUS reboot at any point of time only one node was able to detect the hard disk.
Finally after nearly two months of head scratching (really lost lot of my hair in this :)) we were able to get concurrent logins from both the nodes now! Thanks to the reply by Member no 279089. I bought and installed a TI(TSB43AB23)card on one node and a VIA chipset card on the other. And finally the magic worked. I am able to partition the disk and the partitions are visible from both the nodes.
But now we have a new problem. When we invoke the OCFSTOOL from one node (VIA card node), the previous OCFS partion (the one which was done before the MYSTERIOUS reboot) is still shown. I removed the /etc/ocfs.conf file and still it is being shown in the ocfstool. I tried to ignore this and continued creating the OCFS partition from the other node (TI card node) using the mkfs.ocfs command. That too ran fine without any error messages and was able to mount it under /u02/oradata/orcl. When I try to mount the partition from VIA card node it is giving me the following error
# mount -t ocfs /dev/sdb1 /u02/oradata/orcl
mount: wrong fs type, bad option, bad superblock on /dev/sdb1, or too many mounted file systems
I get the following in /var/log/messages
Jun 29 21:16:40 testrac1 kernel: SCSI disk error : host 2 channel 0 id 0 lun 0 return code = 8000000
Jun 29 21:16:40 testrac1 kernel: Current sd08:11: sense key None
Jun 29 21:16:40 testrac1 kernel: I/O error: dev 08:11, sector 69
I tried googling with the above error message and they seem to point hands towards VIA chipsets. But I am not sure. Please guide us on what needs to be done to resolve this issue if you have faced it before.
Thanks in Advance
Jei

FireWire Disk on RH9

I've found that I have to go through a very precise sequence in order to mount my shared FireWire Hard drive on 3 Red Hat 9.0 nodes (2.4.20-13.9 kernels). Does anyone else have this much difficulty, or can shed some light on my problem? (The node order probably isn't important, it's just the sequence I tested with)
1. Boot node 2, and login.
2. issue the rescan-scsi-bus.sh (drive found)
3. Boot node 3, and login.
4. issue the rescan-scsi-bus.sh (does not find the drive)
5. Boot node 1, and login.
6. issue the rescan-scsi-bus.sh (does not find the drive)
7. On all 3 nodes, issue modprobe -r sbp2
8. On all 3 nodes, issue modprobe sbp2
9. On all 3 nodes, issue rescan-scsi-bus.sh
If I try too much simultaneous disk activity (ex. copying files), it tends to generate errors and abort.

> The default behavior for Firewire disks is that they aren't mounted until a user logs in at the console and are unmounted if no console users are logged in
What makes you say that?
I have several machines with external FireWire drives. Every drive mounts automatically at boot time with no users logged in. I can't think of a single case where this works the way you describe it.

How to boot always from a firewire disk?

Hello,
I hope someone can help me on this.
My HD is almost dead and it's making a lot of noise. I have managed to create a boot disk
on an external disk connected through Firewire.
What shall I do now to have the HD definetly unmounted and have the boot always from the firewire disk?
Thanks in advance
Tom

I've had a similar problem with an external drive that was USB powered.
In my case, the system was looking for a boot device before the drive was ready.
I had a friend that was good with OpenFirmware add a delay before booting to give the drive time to become ready. Added 5 seconds to the boot process, but it works!
Like cldjr said, hold the Option key (⌘) when powering up, if the drive shows up there, I'd guess it's the problem I described...
Anyone good with OpenFirmware??

Firewire disk install on OEL6

I am trying to setup firewire disk with OEL 6 2.6.32-100.34.1.el6uek.i686. I cannot find the rpm for the oracle kernel. Any help, appreciate.

There should be no rpm necessary. Open a terminal and type the following:
sudo modprobe firewire-ohciThis should load the firewire driver and you should be able to access your device.

Firewire disk does'nt boot when updating it from 10.5 to 10.5.1

I installed Leopard for security reasons on an external firewire disk, all went fine with the installation,but when I try to update it to 10.5.1 the disk never boots again ... any solution ????

Encountered the same problem. No solution yet.

External FireWire Disk will not boot

Hi Forum,
after some installations (OS 10.4.9, latest security update and Qucktime 7.1.6 I canot boot from my external disk (Lacie) connected to FireWire 800 port. The Apple shows up but then nothing more happens. Does anyone have a hint?
Thanks
Oliver

Add me to the list also. I had an external firewire disk as a RAID mirror for my boot disk. After a system crash (and some problem with my internal disk) I couldn't boot from the firewire stripe. This used to work.
I've tried creating a single-disk "degraded RAID pair" on the external disk and cloning the internal disk to it (using Disk Utility "Restore"). The cloning works fine but I can't boot from the copy. This is how I originally set up the RAID pair (clone, boot from clone, add internal disk to RAID group). This used to work. I bought a new firewire disk and it has exactly the same problems.
My original firewire disk was an Acomdata E5. My new firewire disk is a Coolmax Gemini CD-311 enclosure with a Hitachi 250GB deskstar inside.
iMac G5 Mac OS X (10.4.9)

Booting from a Firewire Disk backup - get Startup Manager to show FW disk

Hi all,
A hint that I have discovered whilst trying to create a backup on a Firewire external Hard disk and boot my eMac from it.
eMac 1.0GHz (ATI), OS X Tiger 10.4.7 and OWC Mercury Elite Pro 40Gb firewire external disk (Oxford 911+ chipset)
I created a bootable backup using both Carbon Copy Cloner and/or Rsyncx on the firewire disk, but with the same results each time - it woudn't boot from startup manager. I was careful to make sure that the disk was bootable and it would boot from the FW disk after selecting it in system preferences/startup disk and restarting. O.K so far.
OS X also offers an alternative boot method using the startup manager at boot time which you can enter by holding down the option key while turning on the eMac until the startup manager comes up. Normally it shows your "macitoshHD" and "repair" hard drives and you can select the preferred boot up drive. According to Apple help it should also show bootable Firewire disks, but I could not get this to happen.
After a bit of trial and error, I found this sequence reveals the FW dive.
. Make sure your firewire external drive is turned on and plugged in to a firewire port on your eMac.
.Power on the eMac whilst holding down the option key (The key with the apple and squiggley thing next to the spacebar)
.Keep holding the option key until the startup manager icons appear.
.Once the startup manager appears, it will take about 30 secs to 1 min to scan for drives.
. Usually icons showing your internal bootable drives will show along with a rescan and continue icon.
. To get the FW drive to show, hold down the option key and click the rescan button (the turned back arrow icon on the left)
.The system will then rescan and find the firewire drives and show their icons as well as the internal drive icons.
. Select the firewire drive (You did give it a unique name didn't you?) and the click the continue arrow on the right.
. Watch the FW drive activity light and you should see lots of flashing and there you are - booting from the FW drive.
I couldn't find any reference in all the help references to holding down the opt key to rescan, so maybe it is peculiar to this eMac?
Give it a try and be sure to let me know if it works for you.
Ian

OK JMVP, but there is something a little strange here.
After I booted the system whilst holding down the OPTION key, startup manager appeared and scanned for drives but only showed the normal internal drives. (By this time I had already released the OPTION key). I clicked the re-scan icon and same thing, still no FW drive. According to all the Apple help files it should show the external FW drive as well.
But, when I held down the the OPTION key and clicked the rescan icon again, it rescanned and showed the internal drives AND the FW drive. Maybe it's an undocumented feature of startup manager, maybe it's peculiar to the FW drive I'm using
I will try to update the firmware, but the manufacturers website says it doesn't need updating to work with Tiger. Other than this minor prob it already seems to be working OK. Just won't show up in the startup manager without rescanning with the option key pressed.
I did notice whilst searching the help files that there is another option to boot from external drives:
Hold down Option-Command-Shift-Delete keys during startup - this will Bypass primary startup volume and seek a different startup volume (such as a CD or external disk). It didn't work in my case.
Curious?
eMac (ATI) Mac OS X (10.4.7) Mac Newbie

How many firewire disks ???

how many firewire disks can i connect to my mac mini, so it has only one port, and what is the difference between macs equipped with two ports instead of one, there are two different firewire cards or it is just like an internal hub
thanks

The FireWire bus will allow you to connect 67 devices to a single bus. However, I believe the limit of the Finder and the HFS+ (Mac OS Extended) format only allows for 23 different hard drives to appear on your Desktop. You may still be able to mount more, but I'm not entirely sure they'll all be visible on the Desktop...
-Doug
800MHz G4 Cube 1.5GB RAM, 320GB HD Super Drive / 1GHz 12 PBG4 1.25GB RAM & 80GB HD Mac OS X (10.4.3) 1.4GHz Mac mini 1GB RAM, 80GB HD / AirPort Express / EyeTV 200 / 40GB iPod photo

Limiting a user to only 'x' concurrent logins

Hi,
Is there a way of limiting a user to only 'x' concurrent logins.
I've had a hunt around and the only mentions seems to be based on PAM ?
Thanks
al

remote DB admin should be able to log user's activity (I mean ACTUAL user). It does this already even if you use a common DB account - look at "osuser" - you can also get his computer name:
select distinct s.osuser, s.machine, s.username
from sys.v_$session s

Install Solaris on External Firewire Disk

Hello All,
I have a laptop running windows 2000 and I have a External Firewire Hard Disk of 60 GB. Now my question is, " Is it possible to install Solaris 8 on this external Firewire Hard Disk". When I tried installing, It did not find the external Firewire Disk, the only partition information, which solaris was displaying was from my internal Hard Disk.
Please give me clues/information, how should , I do it.
Thanks
Sanjay

First off, is any of your Macs an Intel Mac, that will complicate matters since they use a different hard drive formatting and need a different installation of Mac OS X.
Secondly, you can't directly install Mac OS X from an Intel Mac to an external hard drive, you have to do a clone with a software like Shirt-Pocket's Superduper for that.
If it isn't an Intel Mac, perhaps your Firewire hard drive isn't Mac OS X bootable? Not all are.
Here's a hint on making the Intel/PowerPC partition table:
http://www.macosxhints.com/article.php?story=2006061610374449
Also note, different Macs which came out after Tiger will only boot off the install disk which came with them, and not the retail disk. So you are best off cloning that Mac to its own partition.

Can't open external Firewire disk after upgrade to Leopard

I recently upgraded a PowerBook G4 from 10.4.11(?) to 10.5.5. Now I cannot open its external disk, a LaCie 400 GB Firewire disk, model number 300925, like this: http://www.amazon.com/Lacie-400GB-7200Rpm-Firewire-300925/dp/B00066989Mreport
The disk's icon appears on the desktop after it is plugged in. After double clicking to open it, the pinwheel cursor spins for a while, then Finder appears to restart, redrawing the desktop. When I have the Activity Monitor open during this, I see a process called ReportCrash run for a while before the Finder reset.
Does anyone recognize this symptom? Is there a way to see what is being reported by ReportCrash?

The Genius was skeptical that this was a problem between Leopard and a given disk. Indeed his first test was to boot my laptop from a clean install of Leopard on his own external FireWire drive. Running this way I had no problem opening my external disk from my laptop, running his fresh copy of OSX 10.5.something.
His theory was that something leftover in my configuration files (in /Library/Preferences) from a previous OS was causing the Finder crash while opening a FireWire disk under Leopard. (Since this machine is four years old, there has been a lot of opportunities for kruft to collect.) His suggestion was do a backup (complicated by the fact that I can't open my backup disk, d'oh!) then to do a clean install of Leopard, and then copy back my user folder.
I wish there was a smaller hammer I could use, like setting aside some magic .plist file... Any hints?

Limit number of concurrent logins

Hi, Is there a way to limit the number of concurrent logins in Solaris 10 besides writing a script of some form?
thanks in advance,
Ryan

Oops! I just noticed that in the release notes for 3.6(1) the issues appears as being an open caveat. I am going to upgrade to 3.7 and will post the results.
Josh

Tiger doesn't install on external firewire disk (partition)

hi, i've got a Lacie Porche 250GB firewire drive, partitioned in two parts, and i want to install tiger on the first 50GB partition, but the installer give me an error about the destination: "You cannot install Mac OS X on this volume. Mac OS X cannot start up from this volume."
i'm trying to do this with an ibook G4 and the standard Tiger DVD Install.
the external firewire disk is formatted in HFS+
please help me
iMac Intel\iBook G4 Mac OS X (10.4.8)

A boot disk for the Intel Macs needs to have a GUID Partition Table, the G4 needs the "Apple Partition Map". If you formatted the drive on your Intel Mac, you may need to repartition for the G4. In the "Partition" tab in Disk Utility, use the "Options..." button to choose the partitioning. Backup any data on the drive that you need, first.

Unreliable mounting of Firewire disks and Finder crashing

On both my MacPro3,1 and MacBookPro5,1 running 10.6.2, mounting of Firewire disks is unreliable. Sometimes a partition being mounted does not show up in the Finder (but does appear as mounted to the df command). When that is the case, launching Disk Utility will cause the mounted partition to appear in the Finder. On mounting the disk, the system log always shows these entries:
21/01/2010 14:32:39 diskarbitrationd[14] unable to probe /dev/disk2s1 (status code 0x0000002D).
21/01/2010 14:32:39 diskarbitrationd[14] unable to probe /dev/disk2s2 (status code 0x0000002D).
21/01/2010 14:32:39 diskarbitrationd[14] unable to probe /dev/disk2s3 (status code 0x0000002D).
On dismounting the disk, the system log always shows these entries:
21/01/2010 14:33:13 com.apple.notifyd[13] EV_DELETE failed for file watcher 43
21/01/2010 14:33:13 com.apple.notifyd[13] EV_DELETE failed for file watcher 39
21/01/2010 14:33:13 com.apple.notifyd[13] EV_DELETE failed for file watcher 38
21/01/2010 14:33:13 com.apple.notifyd[13] EV_DELETE failed for file watcher 37
21/01/2010 14:33:13 com.apple.notifyd[13] EV_DELETE failed for file watcher 36
21/01/2010 14:33:13 com.apple.notifyd[13] EV_DELETE failed for file watcher 34
21/01/2010 14:33:13 com.apple.notifyd[13] EV_DELETE failed for file watcher 42
21/01/2010 14:33:13 com.apple.notifyd[13] EV_DELETE failed for file watcher 35
21/01/2010 14:33:13 com.apple.notifyd[13] EV_DELETE failed for file watcher 41
21/01/2010 14:33:14 mdworker[1824] CFPropertyListCreateFromXMLData(): Old-style plist parser: missing semicolon in dictionary.
Sometimes, on dismounting the disk, the Finder crashes, with these entries in the system log:
21/01/2010 14:33:17 com.apple.launchd.peruser.501[165] (com.apple.Finder[175]) Job appears to have crashed: Segmentation fault
21/01/2010 14:33:18 ReportCrash[1827] Saved crash report for Finder[175] version 10.6.2 (10.6.2) to /Users/mike/Library/Logs/DiagnosticReports/Finder2010-01-21-143318Shackleton.crash

Hi Dr. T,
> There is an OnDemand key in there which looks
like it may be doing something useful.
   I'm afraid that I don't know anything about the contents of those files but yes, an OnDemand key sounds like it might do what you want. Although I'm still fuzzy on this point, I gather that what you want is for the device file to be created in /dev but for the volume to not be mounted. As long as the device file is created, you can use the path to the device file as an argument to the "mount" command to mount the drive at any time. I assume that one of "the relevant parameters" in the fstab is the path to the device file.
   I don't know how well I understand what goes on "under the hood" here because I don't do this a lot. I was of the impression that if you eject or detach a drive, the device file is deleted from the device tree. However, if you merely unmount the drive, the device file persists and you can easily mount the drive again.
   By the way, it is supposed to be possible to use fstab with only data that persists between boots. You can use drive UUIDs or labels to refer to partitions. You can use the df command to discover the appropriate device files at the moment and then "diskutil info <device>" to get the UUID.
Gary
~~~~
   If A = B and B = C, then A = C, except where void or
   prohibited by law.
         -- Roy Santoro

Concurrent login firewire disk

Similar Messages

Maybe you are looking for