================================================================================ 3510 password default: login:ras password:agent ================================================================================ Access the Sun StorEdge[TM] 3310, 3510 and 3511 out of band from the command line, use the following: /opt/SUNWsscs/sbin/ #sccli -o -v This will run the se3kxtr out of band and not use tip or in-band. Product Sun StorageTek 3510 FC Array Sun StorageTek 3310 SCSI Array Sun StorageTek 3511 SATA Array ================================================================================ Document Audience: INTERNAL Document ID: 10702191 Title: Post controller upgr and firmware logical drive "incomplete"; Solution: downgrade controller FW Update Date: Tue Jun 28 06:39:58 MDT 2005 Case Number: 10702191 Geo Code (AMER, APAC, EMEA): APAC Synopsis: Post controller upgr and firmware logical drive "incomplete" Product: StorEdge 3510FC HW Platform: Ultra Enterprise 3500 OS Version: Solaris 8 (S8) Engineer: Steve CARRALL Engineer Phone: 02 9844 5307 Open Date: 01-Jun-2005 05:05:08 Close Date: 28-Jun-2005 06:38:29 Resolution downgrade controller FW Problem Customer performed array controller replacement and upgrade to firmware 4.11 on existing SE3510 controllers (370-5537-03), this was successful. As part of FCO #: A0237-2, Sun StorEdge 3510 Raid I/O Module 370-5537-04(Or Less) are effected, SSE replaced 370-5537-03 with latest available from spares (ie. with 4.11 FW) Successfully changed "secondary controller", however after replacing "primary controller" One (ld2) logical drive now has status of "incomplete". When drilling down further through "logical drive" view, two hard disks show "missing". From the main menu the disk drives show status of "used" not the normal "on-line". Solution Several attempts of powercycling array & JBOD were attempted, along with reseting the actual controller. -> none were successful. Idea was delete ld2 & recreate ld2 (without initializing), thus keeping data intact. -> unable to do this on FW 4.11 due to existing bug, but are able to perform this on previous version of FW (3.27R) customer attempted downgrade of FW from 4.11 to 3.27R which was successful. Infact all but one of the previous four "used/missing" drives came back as online. As ld was RAID 5 all data was back. customer then successfully upgraded the FW back to 4.11 original SE3510 (dual controllers) + JBOD .. initial FW to 411 (SES + controller) successful physically replaced controller to get controllers with latest HW rev. Secondary controller replace - all fine then, replaced the original Primary controller - should have failed over to secondary, which it did (note not force failover) - left hand PSU "arc'ed" -> powered off - plugged in new controller, then call LED's on disk on JBOD were ORANGE -> not sure when they went orange (ie. when fibre cables were pulled, or when controller replaced) .. only 10 seconds then went GREEN. - went back into controller thru serial console - logical drives - fatal error -> all drives missing - reset controller -> now ld0, ld1, and ld4 OK .. ld3 "incomplete" Technical Note: advised that bug in new version of FW that prevents deleteling LD & recreating without initializing previous FW version was 3.27R .. is it possible to downgrade from FW 4.11 to 3.27R .. so that we don't hit that bug ?? ... will chase this up Technical Note: As discussed there is a way to try to recover the dead LD with the LD_RESCUE procedure that is still valid for 3.27 f/w revision (but no longer for 4.11...) NOTE, this procedure is UNSUPPORTED & prior to attempting you will need to downgrade the FW from 4.11 to 3.27R. NOTE: you will also need to delete the LD that is currently marked as "incomplete". [START] This procedure should be used if the LD could not be recovered with a controller shutdown and controller reset or a power on reset of the raid unit. This procedure should only be used by qualified and trained personnel. Before you begin, you must have the original LD configuration including ld partions and a prtvtoc of each associated host lun. This includes LD capacity, LD assignment, number of drives, drive channel and ID for each drive in LD, and verify that each drive that was in the missing LD is displayed under View and Edit SCSI Drives with a LG_DRV = NONE and a Status = USED DRV. 1. Remove or fail the secondary controller. 2. Establish serial connection to the remaining controller's text menu. 3. Enter the maintenance mode by typing shift+ZMXNCBV NOTE: This key sequence can be found on solaris system with the following command. " strings | grep OemHotKey " example: bash-2.05$ strings ../113722-03/SUN325S-3310.bin | grep OemHotKey ;OemHotKey="maintenancemode" OemHotKey="ZMXNCBV" 4. type oemmaint at the "maint>" prompt. Note:Possible OEM Maintenance Mode Commands: [q]uit help defaults restore maint fw_version|getserial|set_ld_rescue|clear_ld_rescue 5. type "maint set_ld_rescue" to set parameter. 6. type "q" to return to normal menu. 7. Under View and Edit Logical Drives; on the same LG row # that the missing LD was located, select enter. Note: LG and LD are synonymous 8. Select create LG. 9. Select the SAME drives that were in the missing LD. 10. Select the same capacity per drive. 11. Select the same LD assignment. Note: Check the LD assigment is still correct after Step 17. 12. You'll notice a new menu option called "write LD configuration disabled'. Change this to 'enabled'. 13. Select 'esc'. 14. Select 'yes' to recreate the LD. Note: You'll see the LD progress go immediately to 99% for a short period; this depends on LD size. Then the LD completed message will be displayed. 15. Repeat steps 7 thru 14 for any other missing LDs. 16. Reset the controller to turn LD rescue off and restore the secondary controller. 17. Verify LD assigment. Correct the LD assignment if necessary at this time. 18. Repartion the LDs at this time. 19. Remap the LDs to Host luns. 20. All data should be returned and available ========================================================== Downgrading from Controller Firmware Version 4.11 to 3.xx: ---------------------------------------------------------- If you have a 4.11 replacement controller module for a single-controller configuration, you can choose to downgrade the controller firmware to the 3.xx version. ------- Caution: ------- Do not downgrade from controller version 4.11 to 3.xx unless it is an emergency or you are performing a FRU replacement. If you upgrade an array to 4.11 controller firmware, change one or more parameters, and then downgrade back to 3.xx, you can lose all data. ----- Note: ----- Sun StorEdge Configuration Service, Diagnostic Reporter, and Sun StorEdge CLI version 2.0 are required for all Sun StorEdge 3000 family products. If controller firmware 3.xx arrays are not upgraded to controller firmware 4.11, the new 2.0 software features are not available. Refer to the release notes for your array for a list of new software features. Earlier versions of the controller firmware work with the latest SES, PLD, SAF-TE, and software versions. There is no need to downgrade any of these components if you downgrade the controller version. For instance, CLI version 2.0 is compatible with controller version 3.27. ----- Note: ----- While an upgrade script exists for upgrading from controller firmware version 3.xx to version 4.11, the script is not backward compatible. You cannot use the script to downgrade from controller version 4.11 to version 3.xx. ------- Caution: ------- The CLI download controller command restores factory defaults with downgrades and does not restore: * controller IP address - You must have a serial connection to restore the IP address. And the serial connection must be set to 38400. * net mask and gateway settings * baud rate for the serial port * Host LUN filter settings - type, access mode, and name * channel data rate - FC and SATA arrays: If the channel data rate was set to 1 GHz or 2 GHz, after the download command is run, the channel speed is set to auto (auto-negotiation). Since a FC channel communicates at a rate of either 1 GHz or 2 GHz, the Auto setting uses the best rate for communications at the time. * more than 8 logical drives * customized parameter settings - Record all custom settings prior to the downgrade. The CLI show configuration command does not include all firmware parameters. Be sure to record settings that are in the firmware only, namely sector/head/cylinder parameters and host LUN filter parameters. To downgrade from controller firmware 4.11 to 3.xx firmware, perform the following steps. 1. Change your working directory to the directory in which the patch was unpacked and confirm that the 3.xx controller firmware file is present using the "ls" command. The file must be one of the following: SUN327R-3510.bin for the 3510 FC array (located in patch 113723-07) SUN327R-3511.bin for the 3511 SATA array (located in patch 113724-01) SUN325W-3310.bin for the 3310 SCSI array (located in patch 113722-07) 2. Invoke sccli and verify that you are running sccli version 2.0 with the "version" command. 3. Verify the Product and Revision of the array. Type the command show inquiry Confirm that the correct product name is displayed. Otherwise, this patch does not apply. Select a different device or discontinue installation of this patch. If the firmware version reported is 411G or higher, continue with step 4. 4. Version 3.xx only supports 8 logical drives. To confirm that the array has no more than 8 logical drives, at the "sccli>" prompt, type: show logical-drive -------- Caution: -------- If you have more than 8 logical drives, do not perform this downgrade. You will lose all data above the 8 logical drive limit. Additionally, you might lose or corrupt all data from all logical drives due to conflicting parameters between firmware versions. 5. If you are downgrading due to an emergency and if feasible, save the configuration in a separate location. If you are unable to restore the 3.xx configuration, you can reference this file. At the "sccli>" command prompt, type: show configuration filename.txt and then type: show configuration --xml filename.xml Note: These commands may take several minutes to complete. 6. Stop all I/O to the array before beginning the controller firmware downgrade and unmount any filesystems or volumes mounted from the array. 7. At the "sccli>" prompt, type: download controller-firmware -r filename where filename is: SUN327R-3510.bin for the 3510 FC array SUN327R-3511.bin for the 3511 SATA array SUN325W-3310.bin for the 3310 SCSI array ----- Note: ----- Disregard the CLI message that a script is available to automate the download; the script is only for upgrades. The sccli command will display messages indicating that it is downloading the firmware, programming the controller's flash memory, and "engaging" the new firmware. Wait until the "sccli>" prompt appears again, which may take 10 minutes or more, before proceeding. 8. To re-establish communication with the array, use the serial connection to restore the IP address. netmask, and gateway using the serial interface. 9. Verify that the firmware upgrade succeeded by examining the firmware revision again. At the "sccli>" prompt, type show inquiry and verify that the firmware revision is now reported as "327R" for FC and SATA arrays and "325W" for SCSI arrays. 10. Reconfigure your array to the desired configuration. If you saved a 3.xx configuration file that you want to restore, type download nvram filename where filename is the name of the file that contains the configuration. Otherwise, configure the array using sccli or the firmware application. -------- Caution: -------- Do not restore a 4.11 configuration on a 3.xx controller. This may cause data loss. ======================== Task No: 71744653 Owner: Task Summary: esculate to PTSneed to down rev firmware on 3310 Note: hi, FYI, with the patchadd -d , there is no way we can roll back. Without the option -d, we can rollback by patchrm As for the downgrade of 3310. Please see below action plan. Please bare in mind, there should be a previous backup copy. Just in case we need it. 1. Run extractor for this 3310 eventhough it is running 4.21. Just to be safe. Keep it a side. 2. download the patch 113722-05. This patch will have all the firmware for the controller and safte. (firmware is SUN325S-3310.bin and saftefw_sun2u_1159.bin) 3.we need the sccli to downgrade the firmware. Below method are the same as downgrade. Special Install Instructions: ----------------------------- *** Warning *** Warning *** Warning *** Warning *** Warning *** The RAID controller firmware upgrade procedure will prevent StorEdge 3310 arrays from servicing normal I/O requests for one to three minutes. For this reason, all StorEdge 3310 logical disks should be unmounted from any Solaris hosts connected to the array for the duration of the install procedure. If you are unable to do this, and you have a redundant controller configuration, refer to the manuals for the live upgrade procedure. 1. Stop I/O to the StorEdge 3310 array. Halt applications which may be using the array and, if possible, unmount any filesystems, metadisks, or volume manager volumes mounted from the array. Since the SSCS agent may interfere with firmware downloads, it is necessary to stop any SSCS agent that may be running on any host connected to the array controller. To stop the SSCS agent, log in as root on each host connected to the array and on which the SUNWscsa package has been installed, and issue the shell command "/etc/init.d/ssagent stop". If the SSCS agent is running on a Microsoft Windows system connected to the array, use the "Services" application (Start->Run->"services.msc") to stop the SSCS Agent service. StorADE, or any other application which monitors the status of the enclosure, may also interfere with controller or SAF-TE firmware downloads. Stop the StorADE agent software, or any other software that monitors the array, following the procedures described in the documentation for that application. 2. Change your working directory, if necessary, to the directory in which the patch was unpacked and confirm that the SUN325S-3310.bin file is present using the "ls" command. 3. Invoke the "sccli" command. If only one SE3310 array is connected to the host, it will be selected automatically; otherwise, select the appropriate array from the menu. To upgrade the firmware using an out-of-band connection, specify the hostname or IP address of the array on the sccli command line, or specify the --oob option to use an inband connection to determine the IP address. At the "sccli>" prompt, type "show inquiry" and examine the output. 4. Type the command "show redundancy" at the "sccli>" prompt and examine the output. If the Redundancy mode is not "Active-Active" or the status is not "Enabled", or there is a zero in the Secondary controller serial number field, your configuration cannot be upgraded "live" and you must skip ahead to step 4b. 4a. If the "show redundancy" output shows that the Active-Active mode is Enabled and displays serial numbers for both controllers, then you can perform a live upgrade. Type the command: download controller-firmware SUN325S-3310.bin 5. At the "sccli>" prompt, verify that the new firmware is 325S by typing "show inquiry" and examining the "Revision" field; if the Revision field reports "0325", refer to the first 4 digits of the NVRAM Defaults field. Proceed to step 7 to update the SAF-TE firmware in the same RAID enclosure chassis. 6. At the "sccli>" prompt, verify the SAF-TE firmware version by issuing the "show safte" subcommand: sccli> show safte Ch Id Chassis Vendor Product ID Rev Package --------------------------------------------------------- 0 14 000187 SUN StorEdge 3310 A A000 A000 2 15 005632 SUN StorEdge 3310 D 1159 1159 The "show safte" command will produce one line of output for each enclosure; in the example above, a RAID chassis with serial number (000187) has an expanson chassis (005632) attached, but for a standalone JBOD chassis or RAID subsystems with no expansion chassis attached, only one enclosure will be displayed. Examine the SAFTE firmware revision codes displayed in the "Rev" column, and SAF-TE "Package" revision codes displayed in the "Package" column. In the example above, the first chassis indicates revision A000 and the second indicates revision 1159. In this case, only the first chassis is out of date. If BOTH the Rev column and the Package column contain "1159" or higher for every device reported, AND you are sure that no power supply FRUs have been replaced in any of the enclosures listed, no SAF-TE firmware upgrade is necessary; skip to step 10. Otherwise, continue with step 7. 7. At the "sccli>" prompt, type the following command: download safte-firmware saftefw_sun2u_1159.bin The sccli command will prompt you for confirmation and then proceed with a "live" updated of the SAF-TE firmware. It is not necessary to restart the 3310 RAID controller following the SAF-TE firmware update, but the 3310 RAID controller's telnet and serial interfaces may continue to display the old SAF-TE firmware revision until the next controller reset (sccli will display the correct version number). The SAF-TE firmware download procedure normally takes several minutes to complete. If the selected device is a RAID subsystem with one or more expansion chassis attached, all of the attached expansion chassis will be upgraded at the same time. After this second "download safte-firmware" command has completed, run the sccli subcommand "show safte" again to verify that the SAF-TE "Rev" column contains version "1159" and the "Package" column reports "1159". For example, sccli> show safte Ch Id Chassis Vendor Product ID Rev Package --------------------------------------------------------- 0 14 012345 SUN StorEdge 3310 A 1159 1159 2 15 005632 SUN StorEdge 3310 D 1159 1159 ================================================================================