Solaris 10 Fails To Replace a SCSI Disk under SVM - Application(s) accessing previously attached SATA device have to release it before newly inserted device can be made accessible.

Asset ID:	1-72-2049270.1
Update Date:	2016-08-01
Keywords:

Solution Type Problem Resolution Sure

Solution 2049270.1 : Solaris 10 Fails To Replace a SCSI Disk under SVM - Application(s) accessing previously attached SATA device have to release it before newly inserted device can be made accessible.

Applies to:

Sun Fire X4270 M2 Server - Version All Versions and later
Solaris Operating System - Version 10 3/05 to 10 8/11 U10 [Release 10.0]
Information in this document applies to any platform.

Symptoms

An internal SCSI SATA disk (c1t0d0), a mirrored boot disk, encountered hardware failures and was placed into maintenance state in SVM.

The Device ID of the original failed disk is still listed in metastat output:

Device Relocation Information:
Device   Reloc Device ID
c1t1d0   Yes    id1,sd@SATA_____SEAGATE_ST95000N____________9SP1RZ98
c1t0d0   Yes    id1,sd@SATA_____SEAGATE_ST95000N____________9SP1R57W   <===== Device ID of the original failed disk

The disk is still listed as connected and configured, with condition ok, in cfgadm output:

# cfgadm -al
Ap_Id                          Type         Receptacle   Occupant     Condition
...
sata0/0::dsk/c1t0d0            disk         connected    configured   ok
sata0/1::dsk/c1t1d0            disk         connected    configured   ok

After the disk is unconfigured from SVM (metadetached the submirrors and removed the replicas), unconfigure the disk with cfgadm:

# cfgadm -c unconfigure sata0/0::dsk/c1t0d0

The sata-port is listed as empty and unconfigured in cfgadm output after the disk is physically pulled out:

# cfgadm -al
Ap_Id                          Type         Receptacle   Occupant     Condition
c0                             scsi-bus     connected    configured   unknown
......
sata0/0                        sata-port    empty        unconfigured ok    <===== no disk, as expected
sata0/1::dsk/c1t1d0            disk         connected    configured   ok

Along with the following events in the messages file:

Aug 17 09:31:52 server01 sata: [ID 801593 kern.warning] WARNING: /pci@0,0/pci108e,484c@1f,2:
Aug 17 09:31:52 server01 SATA device detached at port 0
Aug 17 09:31:52 server01 sata: [ID 801593 kern.warning] WARNING: /pci@0,0/pci108e,484c@1f,2:
Aug 17 09:31:52 server01 SATA device detached at port 0

When the new disk is inserted, there are some errors:

Aug 17 09:32:10 server01 sata: [ID 801593 kern.warning] WARNING: /pci@0,0/pci108e,484c@1f,2:
Aug 17 09:32:10 server01 SATA device detected at port 0
Aug 17 09:32:10 server01 sata: [ID 663010 kern.info] /pci@0,0/pci108e,484c@1f,2 :
Aug 17 09:32:10 server01 sata: [ID 761595 kern.info]     SATA disk device at port 0
Aug 17 09:32:10 server01 sata: [ID 846691 kern.info]     model WDC WD500BLHXSUN500G WX71E71XA241 <===== this is a different disk
Aug 17 09:32:10 server01 sata: [ID 693010 kern.info]     firmware     5G08
Aug 17 09:32:10 server01 sata: [ID 163988 kern.info]     serial number WD-WX71E71XA241
Aug 17 09:32:10 server01 sata: [ID 594940 kern.info]     supported features:
Aug 17 09:32:10 server01 sata: [ID 981177 kern.info]      48-bit LBA, DMA, Native Command Queueing, SMART, SMART self-test
Aug 17 09:32:10 server01 sata: [ID 643337 kern.info]     SATA Gen2 signaling speed (3.0Gbps)
Aug 17 09:32:10 server01 sata: [ID 349649 kern.info]     Supported queue depth 32
Aug 17 09:32:10 server01 sata: [ID 349649 kern.info]     capacity = 976773168 sectors
Aug 17 09:32:10 server01 sata: [ID 801593 kern.warning] WARNING: /pci@0,0/pci108e,484c@1f,2:
Aug 17 09:32:10 server01 Application(s) accessing previously attached SATA device have to release it before newly inserted device can be made accessible.    <===== THIS IS THE KEY POINT

The new disk is listed as drive type unknown in format:

     12. c1t0d0 <drive type unknown>
         /pci@0,0/pci108e,484c@1f,2/disk@0,0
     13. c1t1d0 <DEFAULT cyl 60798 alt 2 hd 255 sec 63>
         /pci@0,0/pci108e,484c@1f,2/disk@1,0

The disk is shown as connected and configured in cfgadm, but the condition is listed as unusable:

# cfgadm -al
Ap_Id                          Type         Receptacle   Occupant     Condition
...
sata0/0::dsk/c1t0d0            disk         connected    configured   unusable
sata0/1::dsk/c1t1d0            disk         connected    configured   ok

In more detail from verbose cfgadm output:

# cfgadm -alv
Ap_Id                          Receptacle   Occupant     Condition Information
When         Type         Busy     Phys_Id
...
sata0/0::dsk/c1t0d0            connected    configured   unusable   Mod: WDC WD500BLHXSUN500G WX71E71XA241 FRev: 5G08 SN: WD-WX71E71XA241 <===== new serial number
unavailable disk         n        /devices/pci@0,0/pci108e,484c@1f,2:0
sata0/1::dsk/c1t1d0            connected    configured   ok         Mod: SEAGATE ST95000NSSUN500G 1029M1RZ98 FRev: SF03 SN: 9SP1RZ98
unavailable disk         n        /devices/pci@0,0/pci108e,484c@1f,2:1

Attempting to configure the disk fails or does nothing:

# cfgadm -c configure sata0/0::dsk/c1t0d0
cfgadm: Insufficient condition

# cfgadm -c unconfigure -o unusable_SCSI_LUN sata0/0::dsk/c1t0d0
#

Errors like the following are logged repeatedly in the messages file:

Aug 17 09:33:47 server01 scsi: [ID 107833 kern.warning] WARNING: /pci@0,0/pci108e,484c@1f,2/disk@0,0 (sd0):
Aug 17 09:33:47 server01 Command failed to complete...Device is gone
Aug 17 09:33:47 server01 scsi: [ID 107833 kern.warning] WARNING: /pci@0,0/pci108e,484c@1f,2/disk@0,0 (sd0):
Aug 17 09:33:47 server01 drive offline

Cause

Bug 15601013 : SUNBT6897798-SOLARIS_11 SATA device cannot be configured after disk ejection and reinsertion

Due to devfs checking for the wrong device state.

Solution

When this bug is encountered, the Solaris server must be rebooted in order to recognize the new disk.

Definitive solution is to install a patch with fix and reboot. Fix is contained in the following patches:

Solaris 10 SPARC:

<Patch 148065-02> 02 Feb 2012 Synopsis: Obsoleted by: <Patch 147440-23> SunOS 5.10: devfs patch
<Patch 147440-23> 23 Aug 2012 Synopsis: Obsoleted by: <Patch 147440-24> SunOS 5.10: Solaris kernel patch
<Patch 147440-24> 21 Sep 2012 Synopsis: Obsoleted by: <Patch 147440-25> SunOS 5.10: Solaris kernel patch
<Patch 147440-25> 10 Oct 2012 Synopsis: Obsoleted by: <Patch 147440-26> SunOS 5.10: Solaris kernel patch
<Patch 147440-26> 20 Nov 2012 Synopsis: Obsoleted by: <Patch 147440-27> SunOS 5.10: Solaris kernel patch
<Patch 147440-27> 30 Nov 2012 Synopsis: Obsoleted by: <Patch 147147-26> SunOS 5.10: Solaris kernel patch
<Patch 147147-26> 06 Feb 2013 Synopsis: SunOS 5.10: kernel patch

Solaris 10 x86:

<Patch 148066-02> 02 Feb 2012 Synopsis: Obsoleted by: <Patch 147441-23> SunOS 5.10_x86: devfs patch
<Patch 147441-23> 23 Aug 2012 Synopsis: Obsoleted by: <Patch 147441-24> SunOS 5.10_x86: Solaris kernel patch
<Patch 147441-24> 21 Sep 2012 Synopsis: Obsoleted by: <Patch 147441-25> SunOS 5.10_x86: Solaris kernel patch
<Patch 147441-25> 10 Oct 2012 Synopsis: Obsoleted by: <Patch 147441-26> SunOS 5.10_x86: Solaris kernel patch
<Patch 147441-26> 19 Nov 2012 Synopsis: Obsoleted by: <Patch 147441-27> SunOS 5.10_x86: Solaris kernel patch
<Patch 147441-27> 30 Nov 2012 Synopsis: Obsoleted by: <Patch 147148-26> SunOS 5.10_x86: Solaris kernel patch
<Patch 147148-26> 06 Feb 2013 zip (45.74 MiB) SunOS 5.10_x86: kernel patch

References

<NOTE:1602531.1> - Solaris Volume Manager (SVM) x86 How to Replace a Failed, SCSI Disk, Mirrored with SVM
<NOTE:1482284.1> - Solaris Volume Manager (SVM): How to Boot from a Detached SVM Submirror on a x86 System
<BUG:15601013> - SUNBT6897798-SOLARIS_11 SATA DEVICE CANNOT BE CONFIGURED AFTER DISK EJECTION AND

Attachments

This solution has no attachment