*************************************************** LSI Corp. Nytro MegaRAID Firmware Release for Intel *************************************************** Release date: 2/05/13 ====================== Supported Controllers: ====================== Nytro MegaRAID NMR 8100-4i Nytro MegaRAID NMR 8110-4i Nytro MegaRAID NMR 8120-4i Component: ========= SAS Nytro MegaRAID Firmware Release for Nytro MegaRAID Intel Controllers Release date: 2/05/13 Version Numbers: =============== Current Package Details: ========================= Firmware Package: 23.6.1-0006 Release Date: 2013-02-05 15:30:01 WebBIOS 6.5-8-e_0-Rel Firmware 3.200.15-2275 ROMENV 1.08 PCLI 05.02-01 BootBlock 2.06.00.01-0000 NVDATA 2.1108.03-0089 BootBlockCommon 07.26.05.219 Cache offload Firmware 3.200.14-1825 UEFI_Driver 0x05180000 Hii 1.32.00 FCODE 4.12.05.00 BIOS 5.35.00 Known issues: ============= 1. While migrating from Thunderbolt to Nytro MegaRAID, customers need to delete Nytro Cache volumes created out of external SSDs prior to migration. 2. There are cases where Secure Erase can be aborted by onboard flash drives before it is complete. When that happens, user is expected to reissue Secure Erase again till is succeeds. 3. When user initiates Secure Erase on more than one CC vd, Secure erase completes successfully only on the first VD. User is expected to reissue Secure Erase again for the next VD. 4. On IBM servers with Pcie Gen 3 support, system may not boot with an enclosure connected to the server, when a Nytro Megaraid card is plugged into the Gen 3 slot. Please disconnect the enclosure during boot and connect it back again when server is up and running. 5. Web Bios allows on board flash drives to be configured as Hot spares. This is not an allowed operation, but firmware does not fail this operation at this time. Request user not to try this operation. Errata: ======= 1. Description : Start rebuild option is wrongly displayed in the MSM screen when a Nytro cache drive is in offline state Root cause : Firmware triggers a data flush operation for an offline cache drive, and until flush is complete, it does not accept a rebuild command from the user. Workaround : There is no functional impact because firmware fails the rebuild command during data flush. User is requested to ignore this option or retry until the command succeeds. 2. Description : Apps display the available capacity of supercap as more than its actual capacity. Root cause : This has been identified as a basic MegaRaid issue and has been fixed in Fw package ver 23.7.0-0035. This will be available to Nytro MegaRaid users in subsequent releases. Workaround : The display in the field could be misleading, so request users to ignore it. 3. Description : When user tries to import source VDs with some of them in blocked state (their associated Nytro Cache VD is missing) and some of them in normal state, then he would see all source VDs getting blocked, not only those source VDs which do not have associated Nytro Cache VD. Root cause : Firmware does not maintain a 1:1 mapping of source VD and Nytro Cache VD. Workaround : The user can manually unblock the blocked VDs or bring back the missing Nytro Cache VD, before importing the source VDs 4. Description : The system performance may intermittently dip for 4KB random read and come back to normal on its own Root cause : On-board flash firmware does periodic checks for map compression. Workaround : The recovery is automatic. 5. Description : Unblocking the source VD and rebooting the ESX system may cause Kernel panic Root cause : Unknown Workaround : The user can reboot the system and continue the operation. 6. Description : If a media error is present on unconfigured onboard flash devices and user performs a factory reset operation followed by server reboot , the on-board flash devices will be missing. Root cause : The onboard flash devices will go into a panic state in the presence of media error and would get locked up on reboot Workaround : The user can power down the system, wait for the power discharge to complete and roughly after about 60 seconds power ON the server. Or he has to avoid doing a factory reset before reboot, if media error is detected on the onboard devices 7. Description : The onboard flash FW upgrade for 128 GB Nytro MegaRaid cards will fail on ESX system Root cause : Unknown Workaround : User is requested to upgrade onboard flash fw on non-ESX systems Installation: ============= Use MegaCLI to flash the SAS controllers. MegaCLI can be downloaded from the support and download section of www.lsi.com. Command syntax: MegaCli -adpAllInfo - Provides the index of the Nytro Megaraid card MegaCli -adpfwflash -f ASD_FW_CSA_6MB.rom -an (where n is the index displayed in the previous command) - To upgrade firmware Nytro MegaRAID Behavior: ======================= 1. The user cannot create Nytro Cache volumes with an external SSD. 2. The user cannot create a normal VD using the on-board flash modules. 3. The auto configuration feature does not work when the pinned cache is present. 4. Occasionally, the on-board flash might go into a panic state. When the on-board flash goes into panic state the user gets an event that the drive has been removed, although this is not the only reason for drive to be removed. To recover the drive which is in panic state, the user must shutdown the server, make sure that all the power sources are shutdown(including the super cap), before restarting the server. 5. The maximum size of Nytro Cache volume supported is 740 GB. 6. Only R0 and R1 RAID levels are supported for Nytro Cache VDs 7. Super cap is not hot-pluggable; Removing and inserting with server power on will damage the board & batter functionality 8. Installation of ESXi on source VD is not allowed Cachecade limitations: ====================== Please note the following limitations for firmware based on SSD support, which apply to Nytro flash drives as well. 1. When a user changes the write policy from Write Back to Write Through, the change is not reflected immediately. The time taken to complete this operation is dependent on the amount of dirty data and host I/O at that time. The firmware flushes all the dirty data from the Cachecade virtual drive (VD) to the source. The policy change operation is complete after all the data is flushed. Similarly, deleting a Cachecade VD is also a slow process. 2. The following operations are not supported on Cachecade VD: • Copy back • Patrol Read • Check Consistency • Initialization • Reconstruction 3. SSD caching should not be enabled on a source VD created using the SSD drives. 4. The application does not report any progress during the delete Cachecade VD operation. 5. When a cachecade VD is deleted and system rebooted before delete operation is complete, user might have to re-issue the delete operation after the system is up, because firmware does not remember the state of the cachecade VD after reboot 6. When a rebuild operation is initiated on a Cachecade VD, the rebuild does not start immediately. The time taken to start the rebuild operation is dependent on the amount of dirty data and host I/O at that point of time. The firmware flushes all the dirty data from the Cachecade VD to the source.