update to 2.6.18-92.1.22.el5 and firewire disk not found

Installation and support for Oracle DB on CentOS.
kmcq
Posts: 8
Joined: 2009/03/17 00:59:20

update to 2.6.18-92.1.22.el5 and firewire disk not found

Post by kmcq » 2009/03/17 01:14:14

I had installed Centos 5.1 2.6.18-92.1.13 for RAC testing using a modification of Jeff Hunter's Article, Cluster had been working fine for months. Just recently updated and now the Firewire disk is not recognized.

Looked at some of the other posts and no help.

I am trying to stay as vanilla as possible, so kernel is NOT centos-plus as other post have had issues with.

My question now is what can I do ?

Option 1 -

Revert to 5.1 which may require re-install

Option 2 -

If there is there a go forward path for 5.2 get this working

Option 3 -

Wait for 5.3 if this fixes the issue

In my research found that the ieee1394 driver was updated and the release notes for 5.3 suggest that it was broken in 5.2.

I apologize if this is not the right forum for hardware issues, but since this configuration was driven by Oracle RAC requirements thought that this group would have similar issues and concerns.


If nothing more than to warn others that 5.1 works and 5.2 does not for firewire RAC config.

:-(

This is a test box so, not an urgent issue.

Kevin

gerald_clark
Posts: 10642
Joined: 2005/08/05 15:19:54
Location: Northern Illinois, USA

Re: update to 2.6.18-92.1.22.el5 and firewire disk not found

Post by gerald_clark » 2009/03/17 13:30:54

Can't you just boot the old kernel?

kmcq
Posts: 8
Joined: 2009/03/17 00:59:20

Re: update to 2.6.18-92.1.22.el5 and firewire disk not found

Post by kmcq » 2009/03/17 19:05:15

Tried going back on one box, but may have corrupted thing.

Boxes are Adam/Eve

ADAM is go forward box. EVE is still pretty much untouched.

Question is how far back and what else has changed.

from grup booting EVE with -13 kernel still does not find firewire disk
:-?

gerald_clark
Posts: 10642
Joined: 2005/08/05 15:19:54
Location: Northern Illinois, USA

Re: update to 2.6.18-92.1.22.el5 and firewire disk not found

Post by gerald_clark » 2009/03/17 19:31:38

Boot the kernel you were running before the upgrade.

kmcq
Posts: 8
Joined: 2009/03/17 00:59:20

Re: update to 2.6.18-92.1.22.el5 and firewire disk not found

Post by kmcq » 2009/03/19 06:14:19

More Info

Adam - Box Installed OLE 5.3 - No luck

Upgraded Adam to 2.6.18-128.0.0.0.2 - Oracle Patch - No luck

Upgraded to -128.1 using http://centos.toracat.org/ajb/tmp/kernels/ - No Luck

Went back and Installed Centos from 5.1 DVD

Upgraded all modules , ended up with 2.6.18-92.1.22

Installed: ieee1394-kmdl-2.6.18-92.1.22.el5 - 2.6.18-2.el5.i686 from ATRPMS

Back in Business - Sort of


[root@adam ~]# find / -name "*ieee1394*" | grep modules
/lib/modules/2.6.18-53.el5/modules.ieee1394map
/lib/modules/2.6.18-92.1.22.el5/modules.ieee1394map
/lib/modules/2.6.18-92.1.22.el5/updates/drivers/ieee1394
/lib/modules/2.6.18-92.1.22.el5/updates/drivers/ieee1394/ieee1394.ko
[root@adam ~]# md5sum /lib/modules/2.6.18-92.1.22.el5/updates/drivers/ieee1394/ieee1394.ko
d4c1ae586889aadff418f6bbf6734264 /lib/modules/2.6.18-92.1.22.el5/updates/drivers/ieee1394/ieee1394.ko


root@adam ~]# grep -i ieee /var/log/messages | tail
Mar 18 22:59:39 adam kernel: ieee1394: sbp2: Maximum concurrent logins supported: 2
Mar 18 22:59:39 adam kernel: ieee1394: sbp2: Number of active logins: 0
Mar 18 22:59:39 adam kernel: ieee1394: sbp2: Logged into SBP-2 device


Mar 18 23:01:59 adam kernel: ieee1394: Error parsing configrom for node 0-00:1023
Mar 18 23:01:59 adam kernel: ieee1394: Error parsing configrom for node 0-01:1023
Mar 18 23:02:00 adam kernel: ieee1394: Error parsing configrom for node 0-02:1023
Mar 18 23:02:01 adam kernel: scsi11 : SBP-2 IEEE-1394
Mar 18 23:02:02 adam kernel: ieee1394: sbp2: Maximum concurrent logins supported: 2
Mar 18 23:02:02 adam kernel: ieee1394: sbp2: Number of active logins: 1
Mar 18 23:02:02 adam kernel: ieee1394: sbp2: Error logging into SBP-2 device - login failed

[root@eve ~]# grep -i ieee /var/log/messages | tail
Mar 18 18:49:14 eve yum: Erased: ieee1394-kmdl-2.6.18-92.1.13.el5
Mar 18 18:49:18 eve yum: Erased: ieee1394-kmdl-2.6.18-92.1.22.el5
Mar 18 18:50:13 eve yum: Installed: ieee1394-kmdl-2.6.18-92.1.22.el5 - 2.6.18-2.el5.i686
Mar 18 18:53:32 eve kernel: scsi4 : SBP-2 IEEE-1394
Mar 18 18:53:32 eve kernel: ieee1394: sbp2: Driver forced to serialize I/O (serialize_io=1)
Mar 18 18:53:32 eve kernel: ieee1394: sbp2: Try serialize_io=0 for better performance
Mar 18 22:50:08 eve kernel: scsi4 : SBP-2 IEEE-1394
Mar 18 22:50:08 eve kernel: ieee1394: sbp2: Driver forced to serialize I/O (serialize_io=1)
Mar 18 22:50:08 eve kernel: ieee1394: sbp2: Try serialize_io=0 for better performance
Mar 18 23:02:07 eve kernel: scsi5 : SBP-2 IEEE-1394


Eve can mount the disk, but steals it from Adam - Multi-host driver not installed.

Will need be re-installing on Eve.


by the way

[root@adam ~]# find / -name "*ieee1394*" | grep modules
/lib/modules/2.6.18-53.el5/modules.ieee1394map
/lib/modules/2.6.18-92.1.22.el5/modules.ieee1394map
/lib/modules/2.6.18-92.1.22.el5/updates/drivers/ieee1394
/lib/modules/2.6.18-92.1.22.el5/updates/drivers/ieee1394/ieee1394.ko

Clean

v.s.

[root@eve ~]# find / -name "*ieee1394*" | grep modules
/lib/modules/2.6.18-92.1.22.el5/updates/drivers/ieee1394
/lib/modules/2.6.18-92.1.22.el5/updates/drivers/ieee1394/ieee1394.ko
/lib/modules/2.6.18-92.1.22.el5/weak-updates/ieee1394
/lib/modules/2.6.18-92.1.22.el5/weak-updates/ieee1394/ieee1394.ko
/lib/modules/2.6.18-92.1.22.el5/modules.ieee1394map
/lib/modules/2.6.18-53.el5/weak-updates/ieee1394
/lib/modules/2.6.18-53.el5/weak-updates/ieee1394/ieee1394.ko
/lib/modules/2.6.18-53.el5/modules.ieee1394map
/lib/modules/2.6.18-92.1.13.el5/weak-updates/ieee1394
/lib/modules/2.6.18-92.1.13.el5/weak-updates/ieee1394/ieee1394.ko
/lib/modules/2.6.18-92.1.13.el5/modules.ieee1394map
/lib/modules/2.6.18-92.el5/extra/ieee1394
/lib/modules/2.6.18-92.el5/extra/ieee1394/ieee1394.ko

Lots of extra stuff



Don't know if anyone else cares, but there are few references on web for this firewire config since Jeff Hunter moved to iSCSI for RAC.


The firewire technology works, but not hard to maintain. Sigh ...

User avatar
toracat
Site Admin
Posts: 7518
Joined: 2006/09/03 16:37:24
Location: California, US
Contact:

Re: update to 2.6.18-92.1.22.el5 and firewire disk not found

Post by toracat » 2009/03/19 10:00:38

From the last set of output, it looks like a kernel-independent version of the ieee1394 module has been installed on Eve but not on Adam. Can you run the following command on Eve and show us the output?

ls -l `find /lib/modules -name ieee1394.ko`

You will find symbolic links there. Also I'm curious to see the output of:

rpm -qf /lib/modules/2.6.18-92.1.22.el5/updates/drivers/ieee1394/ieee1394.ko

rpm -qf /lib/modules/2.6.18-92.el5/extra/ieee1394/ieee1394.ko

kmcq
Posts: 8
Joined: 2009/03/17 00:59:20

Re: update to 2.6.18-92.1.22.el5 and firewire disk not found

Post by kmcq » 2009/03/19 19:25:55

[root@eve ~]# ls -l `find /lib/modules -name ieee1394.ko`
lrwxrwxrwx 1 root root 53 Mar 15 12:01 /lib/modules/2.6.18-53.el5/weak-updates/ieee1394/ieee1394.ko -> /lib/modules/2.6. 18-92.el5/extra/ieee1394/ieee1394.ko
lrwxrwxrwx 1 root root 53 Mar 15 12:01 /lib/modules/2.6.18-92.1.13.el5/weak-updates/ieee1394/ieee1394.ko -> /lib/modules /2.6.18-92.el5/extra/ieee1394/ieee1394.ko
-rw-r--r-- 1 root root 105636 Feb 15 20:36 /lib/modules/2.6.18-92.1.22.el5/updates/drivers/ieee1394/ieee1394.ko
lrwxrwxrwx 1 root root 53 Mar 15 12:01 /lib/modules/2.6.18-92.1.22.el5/weak-updates/ieee1394/ieee1394.ko -> /lib/modules /2.6.18-92.el5/extra/ieee1394/ieee1394.ko
-rw-r--r-- 1 root root 1116985 Jan 29 14:41 /lib/modules/2.6.18-92.el5/extra/ieee1394/ieee1394.ko


[root@adam log]# ls -l `find /lib/modules -name ieee1394.ko`
-rw-r--r-- 1 root root 105636 Feb 15 20:36 /lib/modules/2.6.18-92.1.22.el5/updates/drivers/ieee1394/ieee1394.ko


[root@eve ~]# rpm -qf /lib/modules/2.6.18-92.1.22.el5/updates/drivers/ieee1394/ieee1394.ko
ieee1394-kmdl-2.6.18-92.1.22.el5-2.6.18-2.el5
[root@eve ~]# rpm -qf /lib/modules/2.6.18-92.el5/extra/ieee1394/ieee1394.ko
kmod-ieee1394-1.0.0-1.el5



[root@adam log]# rpm -qf /lib/modules/2.6.18-92.1.22.el5/updates/drivers/ieee1394/ieee1394.ko
ieee1394-kmdl-2.6.18-92.1.22.el5-2.6.18-2.el5
[root@adam log]# rpm -qf /lib/modules/2.6.18-92.el5/extra/ieee1394/ieee1394.ko
error: file /lib/modules/2.6.18-92.el5/extra/ieee1394/ieee1394.ko: No such file or directory


Also, checking logs on EVE found the following for first re-boot after upgrade

Mar 13 20:10:21 eve kernel: ieee1394: no version for "struct_module" found: kernel tainted.

from yum log - the update of kernel

Mar 13 16:24:29 Installed: kernel - 2.6.18-92.1.22.el5.i686

Attempt to fix

Mar 15 12:02:19 Installed: kmod-ieee1394 - 1.0.0-1.el5.i686
Mar 15 12:06:02 Installed: ieee1394-kmdl-2.6.18-92.1.22.el5 - 2.6.18-2.el5.i686

messages on first reboot after modules installed

Mar 15 12:16:22 eve kernel: firewire_core: BM lock failed, making local node (ffc0) root.
Mar 15 12:16:28 eve kernel: firewire_core: Unsolicited response (source ffc0, tlabel 19)
Mar 15 12:16:28 eve kernel: firewire_core: Unsolicited response (source ffc0, tlabel 1a)
Mar 15 12:16:29 eve kernel: firewire_core: created new fw device fw1 (2 config rom retries, S400)
Mar 15 12:16:29 eve kernel: scsi4 : SBP-2 IEEE-1394
Mar 15 12:16:29 eve kernel: firewire_sbp2: logged in to fw1.0 LUN 0000 (0 retries)
Mar 15 12:16:29 eve kernel: ieee1394: sbp2: Driver forced to serialize I/O (serialize_io=1)
Mar 15 12:16:29 eve kernel: ieee1394: sbp2: Try serialize_io=0 for better performance
Mar 15 12:16:34 eve kernel: firewire_sbp2: sbp2_scsi_abort
Mar 15 12:16:44 eve kernel: firewire_sbp2: sbp2_scsi_abort
Mar 15 12:16:44 eve kernel: scsi 4:0:0:0: scsi: Device offlined - not ready after error recovery
Mar 15 12:16:48 eve kernel: firewire_sbp2: released fw1.0
Mar 15 12:16:58 eve kernel: scsi5 : SBP-2 IEEE-1394
Mar 15 12:16:58 eve kernel: firewire_core: created new fw device fw1 (0 config rom retries, S400)
Mar 15 12:16:58 eve kernel: firewire_core: created new fw device fw2 (0 config rom retries, S400)
Mar 15 12:16:58 eve kernel: firewire_sbp2: error status: 0:4
Mar 15 12:16:59 eve last message repeated 4 times
Mar 15 12:16:59 eve kernel: firewire_sbp2: logged in to fw1.0 LUN 0000 (4 retries)
Mar 15 12:17:05 eve kernel: firewire_sbp2: sbp2_scsi_abort
Mar 15 12:17:15 eve kernel: firewire_sbp2: sbp2_scsi_abort
Mar 15 12:17:15 eve kernel: scsi 5:0:0:0: scsi: Device offlined - not ready after error recovery
Mar 15 12:17:15 eve logger: Cluster Ready Services waiting on dependencies. Diagnostics in /tmp/crsctl.4145.
Mar 15 12:17:15 eve logger: Cluster Ready Services waiting on dependencies. Diagnostics in /tmp/crsctl.4161.
Mar 15 12:17:15 eve logger: Cluster Ready Services waiting on dependencies. Diagnostics in /tmp/crsctl.4288.
Mar 15 12:18:10 eve kernel: firewire_sbp2: released fw1.0


...

Will keep EVE as is for now.

kmcq
Posts: 8
Joined: 2009/03/17 00:59:20

Re: update to 2.6.18-92.1.22.el5 and firewire disk not found

Post by kmcq » 2009/03/19 19:48:31

Btw, Forgot to say thank you for interest and support :-)

User avatar
toracat
Site Admin
Posts: 7518
Joined: 2006/09/03 16:37:24
Location: California, US
Contact:

Re: update to 2.6.18-92.1.22.el5 and firewire disk not found

Post by toracat » 2009/03/19 21:26:29

I am slowly beginning to understand your situation after seeing the output. You have installed:

Mar 15 12:02:19 Installed: kmod-ieee1394 - 1.0.0-1.el5.i686
Mar 15 12:06:02 Installed: ieee1394-kmdl-2.6.18-92.1.22.el5 - 2.6.18-2.el5.i686

on [b]eve[/b]. But [b]adam[/b] has only the second one (kmdl). If firewire is working on [b]eve[/b], but not on [b]adam[/b] (is this correct?), then chances are that the driver from the first (kmod-ieee1394) is doing the job on [b]eve[/b]. I believe you have downloaded this from [b]Alan[/b]'s collection?

In any event, let's find out which driver is in use by issuing a command:

/sbin/modinfo ieee1394

on both machines. I also want to see the running kernel on both:

uname -mr

kmcq
Posts: 8
Joined: 2009/03/17 00:59:20

Re: update to 2.6.18-92.1.22.el5 and firewire disk not found

Post by kmcq » 2009/03/19 22:51:08

[root@adam log]# /sbin/modinfo ieee1394
filename: /lib/modules/2.6.18-92.1.22.el5/updates/drivers/ieee1394/ieee1394.ko
license: GPL
srcversion: 9193120996DAD27256F2DBC
depends:
vermagic: 2.6.18-92.1.22.el5 SMP mod_unload 686 REGPARM 4KSTACKS gcc-4.1
parm: ignore_drivers:Disable automatic probing for drivers. (int)
parm: fcp:Map FCP registers (default = 1, disable = 0). (int)
parm: disable_nodemgr:Disable nodemgr functionality. (int)
parm: disable_irm:Disable Isochronous Resource Manager functionality. (bool)
[root@adam log]# md5sum /lib/modules/2.6.18-92.1.22.el5/updates/drivers/ieee1394/ieee1394.ko
d4c1ae586889aadff418f6bbf6734264 /lib/modules/2.6.18-92.1.22.el5/updates/drivers/ieee1394/ieee1394.ko
[root@adam log]# uname -mr
2.6.18-92.1.22.el5 i686


[root@eve log]# /sbin/modinfo ieee1394
filename: /lib/modules/2.6.18-92.1.22.el5/weak-updates/ieee1394/ieee1394.ko
license: GPL
srcversion: 9193120996DAD27256F2DBC
depends:
vermagic: 2.6.18-92.el5 SMP mod_unload 686 REGPARM 4KSTACKS gcc-4.1
parm: ignore_drivers:Disable automatic probing for drivers. (int)
parm: fcp:Map FCP registers (default = 1, disable = 0). (int)
parm: disable_nodemgr:Disable nodemgr functionality. (int)
parm: disable_irm:Disable Isochronous Resource Manager functionality. (bool)
[root@eve log]# md5sum /lib/modules/2.6.18-92.1.22.el5/weak-updates/ieee1394/ieee1394.ko
aff5eb215df03f92e2978fa0805a4f48 /lib/modules/2.6.18-92.1.22.el5/weak-updates/ieee1394/ieee1394.ko
[root@eve log]# uname -mr
2.6.18-92.1.22.el5 i686


and it is ADAM that is working correctly now not EVE

Checking checking checking

thought that I had erased kmod-ieee1394 - 1.0.0-1.el5.i686 from EVE but had not.

did

rpm -e kmod-ieee1394

restarted - no luck

checked blacklist-firewire

had commented out the firewire line - removed comment

init 6 -

EVE is now working as well as ADAM

:-D

Post Reply