LSI 1068E Ctrl+M Software RAID: Configuration, Spares and Recovery Examples
This archived guide is labeled LSI 1068E by the source and shows the LSI Software RAID Configuration Utility. Screens mix version A.53 (September 6, 2007) and A.60 (July 30, 2008), different disks and different capacities. Retain those as separate examples. Confirm the actual controller, firmware, supported RAID key and driver before using a configuration path.
Back up required data and record the actual member identities before configuration or recovery. Deleting a volume, clearing configuration, choosing New Configuration, rebuilding the wrong destination or forcing a stale member online can destroy access or overwrite current data. Online or a saved configuration does not establish data consistency, completed initialization or OS/application recovery.
1. Enter the configuration utility
At the SAS/SATA controller POST prompt, press Ctrl+M. The source POST picture detects the RAID key and lists three MAX3147RC disks. The next picture shows the utility Management Menu:


2. View a virtual drive
Choose Objects → Virtual Drive, select the intended VD and press Enter. Open View/Update Parameters to inspect level, size, strip size, member/stripe count, state, spanning and cache settings:



The parameter example is RAID1, 304222 MB, 64 KB strips, two stripes and one span, with Disk WC Off and Read Ahead On. DWC controls disk write caching; RA controls read-ahead. The illustrated DWC menu can change its state, but changing it is a separate policy decision, not a required recovery operation:

3. Delete configuration only when intended
To delete one VD, use Objects → Virtual Drive, highlight the verified target, press Del and review Want To Delete A VD? Yes confirms deletion; the captured dialog is still on No. The source treats this as destructive and requires a backup beforehand:


Configure → Clear Configuration removes the controller’s current configuration as a whole. Review its scope before Yes; the picture again shows the No choice highlighted. Do not use Clear Configuration to troubleshoot an array containing required data:


4. Configure RAID0
Choose Configure, then Easy Configuration or the applicable View/Add Configuration path. New Configuration replaces the existing configuration when saved; it is a different, destructive choice:


Use Space to mark the intended available disks. Their labels change from READY to ONLIN Axx-xx as they are assigned; this indicates configuration membership rather than completed disk testing. Press F10, use Space in the configurable-array list to select the intended array/span, then F10 to configure its VD:




Open RAID and select RAID0. Review size, DWC and RA, then Accept. The source’s later RAID0 form reads 608444 MB, DWC Off, RA On and SPAN YES, while a background/intermediate list reads RAID0 1216888 MB. Preserve the actual form/result distinction rather than deriving a new universal capacity rule from these mixed screens:



Exit the configuration form with Esc and answer the intended Save Configuration prompt. The later saved example lists RAID0 608444 MB, four stripes, 64 KB, ONLINE. It confirms the source configuration save, without separate initialization, filesystem or boot tests:



5. Configure RAID1
Use Configure → Easy Configuration or View/Add Configuration as appropriate; keep the New Configuration warning in mind. Space selects the intended two members, then F10 opens array selection. Select the array and press F10 again:





Select RAID1. The illustrated form uses 304222 MB, DWC Off, RA On and SPAN NO. Accept, exit with Esc and save the verified configuration. The parameter result shows RAID1, 304222 MB, 64 KB, two stripes, one span and OPTIMAL:





6. Configure RAID10
Follow the same configuration-method path. The source requires an even member count of at least four and illustrates four selected WDC WD3200YS drives. Use Space to select the intended members and F10 to proceed, then select the intended array/span and press F10:






The RAID menu includes RAID10. Review the final RAID10 form rather than a preceding RAID0 form. The source shows 608444 MB, DWC Off, RA On and SPAN YES. Accept, exit with Esc, save, and review the saved VD list, which shows RAID10 608444 MB, four stripes, 64 KB and ONLINE:






7. RAID5: distinguish the described option from the evidence
The source repeats Configure → Easy Configuration or View/Add Configuration, physical-disk selection and the F10 array-selection steps. Its first images in this branch include an already saved RAID0, rather than an empty configuration:






The source says RAID5 requires the appropriate RAID key and is unavailable/gray without activation. It also says its supported RAID5 configuration fixes the strip size at 64 KB. Check that capability on the exact controller and firmware; do not infer support from a generic MegaRAID manual. In the displayed level menu RAID5 is gray and the active form remains RAID0:

Although the source prose concludes that RAID5 has been created, its last picture shows assigned physical drives, not a RAID5 VD result. The provided images therefore do not verify successful RAID5 creation. If the required licensed option is available on the actual supported controller, use its documented Accept/Save process and verify the resulting VD level separately:

8. Configure a hot spare
The source illustrates a RAID1 case with two members and remaining READY drives. Under Configure → Easy Configuration, highlight the verified spare candidate and press F4. Confirm HotSpare? only for the intended available disk:


A usable spare must meet the controller’s capacity and compatibility requirements; an undersized-disk warning is not permission to use an unsuitable spare. The source text mentions such a warning, but the provided dialog simply asks HotSpare?. The next picture is the actual HOTSP state:

A spare remains available while redundant arrays are healthy and may be used for rebuilding an eligible failed redundant array. It cannot restore RAID0 redundancy after a member failure. The source mentions Force Offline to remove a spare designation; use that action only on the verified spare, not an active member, and follow the exact controller’s removal procedure. Esc returns to the save prompt:

9. A redundant-array member goes offline
The next example changes utility version and drive type: it shows RAID1 285148 MB, 64 KB, State DEGRADED, Disk WC Off and Read Ahead On. These are not the preceding 304222 MB WDC-drive case:

Protect readable required data before proceeding. With a suitable spare, the controller may start rebuilding automatically. Without one, follow the supported replacement/rebuild procedure. The source illustrates a disk it says is not physically damaged: in Rebuild, verify the intended destination, use Space to select its FAIL entry and F10 to start the requested rebuild. Do not write to a surviving/current member merely because another slot label looks similar:

Confirm the intended rebuild with Y. The source’s progress capture then shows Drive 1 rebuilding at 0% Completed. It demonstrates a started operation, not successful completion; verify the actual completed rebuild and data access afterward:


10. Offline arrays and the source Force Online example
The source discusses multiple offline members and proposes forcing the last failed member online first. When failure order is unknown, it further proposes trying one member, forcing it offline again if the array does not recover, then trying another. These are the source’s recovery suggestions, not a verified or generally safe procedure for choosing authoritative data. For required data, preserve the failure history and member identities and obtain qualified recovery support before state changes or rebuild writes.
The supplied pictures themselves show a two-member list with one ONLIN and one FAIL entry, not a documented multi-failure history. Under Objects → Physical Drive the menu offers Force Online. The next dialog asks Change To Online?; Yes changes the selected disk’s controller state. The last picture shows both entries ONLIN. That state change alone does not establish correct failure ordering, data freshness, a healthy volume or successful recovery:



The source recommends backing up data if a recovered system becomes readable, then addressing the remaining failed member through the single-member procedure. A successful boot alone does not validate all data. Do not initialize, clear configuration or perform a trial rebuild when the authoritative members are uncertain.
References
The LSI English Configuration Software guide documents the related Ctrl+M menu family and hot-spare rules. It is used to check concepts and control meanings, not to replace the source’s older software-RAID firmware capabilities or screenshots.