LSI MegaRAID SAS 9260-8i Data Recovery: 9260-8i, 9260-4i, 9260CV and 9260DE Failures, Error Messages, and Replacement Traps

The LSI MegaRAID SAS 9260-8i was the workhorse of the first 6 Gb/s MegaRAID generation: eight internal ports, 512 MB of cache and an optional battery on a low-profile card that went into white-box servers, storage builds and workstations from 2010 on — and, rebadged as the IBM ServeRAID M5015 and M5014, into a generation of System x servers. Many are still running. When one fails today, it is usually in a server a decade old, with drives of the same age, and the problems are specific: drives that drop in groups, replacement cards that will not import, batteries long past their service life, and cards that overheat outside the rack they were designed for.

This page covers what is specific to the 9260-8i and its close siblings. The full catalog of SAS2108 messages — POST screens, WebBIOS states, MSM events, MegaCLI commands and Linux driver messages, each rated by risk — is on our LSI MegaRAID SAS2108 recovery page.

The 9260 Family

	9260-8i	9260-4i	9260CV-8i / CV-4i	9260DE-8i
Internal ports	8 (2 × SFF-8087, vertical)	4 (1 × SFF-8087)	8 / 4	8 (same connectors as 9260-8i)
Cache	512 MB DDR2-800	512 MB	512 MB non-volatile	512 MB
Cache protection	Optional iBBU07 battery	Optional battery	CacheVault flash module (CVFM01)	—
Chip / host	LSI SAS2108 (800 MHz PowerPC), PCIe 2.0 x8, low-profile MD2

The 9261-8i is a close cousin on the same chip with the same 512 MB cache; its two connectors exit horizontally instead of vertically. The IBM ServeRAID M5015 (512 MB, battery included) and M5014 (256 MB, battery optional) are the same chip with IBM firmware, and IBM requires its M5000 Advanced Feature Key for RAID 6 and 60.

Drives That Drop in Groups

The single most common 9260-8i story is not a drive failure. It is several drives vanishing together and coming back as (Foreign) Unconfigured Bad, with the virtual drive offline.

Two documented cases show both sides of it. In one, an administrator found four of eight drives in a 9260-8i RAID 6 marked failed, all behind an HP SAS expander. Moving the drives to another backplane made them look healthy, and LSI support walked him through making each drive Unconfigured Good and rescanning; the array came back. In the other, an IBM M5014 — the 9260-8i’s twin — running drives on SFF-8087-to-SATA breakout cables kept losing an entire RAID 10 every couple of weeks, with every member marked foreign each time, and the owner kept recovering it with the same make-good-and-import routine.

Each connector on a 9260-8i carries four drives. When several drives drop at the same moment, especially all four on one connector or everything behind one expander, suspect the cable, connector, backplane, expander or power supply before the drives. That is also the one situation where the vendor’s make-good-and-import procedure is usually safe. What makes it dangerous is repetition: every cycle is a chance for a real drive failure to happen in between, and then the next import trusts a drive that no longer holds current data.

If your 9260-8i has done this more than once, back up the array before touching anything else, and fix the path — cable, backplane, expander, power — before the next event.

The Identical Replacement That Will Not Import

When a 9260-8i dies, the obvious fix is another 9260-8i. It usually works. When it does not, the controller shows the foreign configuration at POST but refuses to import it, and MegaRAID Storage Manager reports Foreign drives cannot be imported as they don’t form a complete configuration.

One administrator documented exactly this after swapping a failed card for an identical model; the forum’s first advice was to flash the replacement to the same firmware as the original. Firmware 12.12.0-0111’s release notes (January 2013) show how much POST and import behavior can change between releases. The owner eventually cleared the configuration and recreated his RAID 1 without initialization, and it booted — while noting that clearing and recreating will not always get data back.

The safe order on a 9260-8i replacement:

1. Install the replacement with no drives attached and check its firmware version.
2. Match the original card’s firmware if you know it, with the drives still disconnected.
3. Connect the drives and look at the foreign configuration preview before importing anything.
4. If the import is refused, stop. Clearing and recreating is a one-shot gamble, and on RAID 5, 6, 50 or 60 the odds are poor.

IBM Cards Cross-Flashed to LSI Firmware

Because the IBM M5014 and M5015 use the same chip, they are commonly cross-flashed to LSI 9260 firmware in home labs and budget servers. That is a reasonable thing to do with a new card and no drives attached. With array members attached, a firmware change can leave the controller unable to deliver its cache — MegaRAID has a specific fatal event for cache discarded due to firmware version incompatibility — and a cross-flashed card may not read an array the IBM firmware wrote. If an array is already in trouble, flash nothing until the data is safe.

Heat Outside the Rack

The 9260-8i is a server card that needs active airflow across its heatsink. In a tower, a desktop case or a storage chassis with a failed fan, it runs hot. An overheating controller does not usually announce itself; it shows up as command timeouts, controller resets in the operating system log (on Linux, megasas: RESET lines and, in the worst case, FW detected to be in faultstate, restarting it…), and drives dropping out of arrays. Check airflow before blaming the drives, and back up before opening the case.

The iBBU07 Battery: Old, Learning, or Gone

The 9260-8i’s optional iBBU07 battery mounts directly on the card and is rated to hold cache for up to 72 hours. Most batteries on cards still in service are years past their expected life. Three things follow:

• Dead or removed battery with write-back forced on: a power loss drops the cache, and the next boot may show Memory/battery problems were detected. The adapter has recovered, but cached data was lost. The event log records Controller cache discarded due to memory/battery problems. File systems may need careful checking; image before running repair tools.
• Learn cycles: the battery’s periodic calibration changes cache behavior while it runs. Firmware 12.12.0-0111 lists a fix for the controller quiescing I/O when a learn cycle started even in “write-back always” mode — a busy array stalling on schedule.
• Outages longer than the battery: after about three days without power, cached writes are gone regardless.

On the 9260CV models, a CacheVault flash module replaces the battery and keeps the cache contents in non-volatile memory, so extended outages are less of a concern.

9260-8i Messages: Quick Index

• “Foreign configuration(s) found on adapter… ‘F’ to import”
• “All of the disks from your previous configuration are gone”
• “There are offline or missing virtual drives with preserved cache”
• “Memory/battery problems were detected…”
• “(FOREIGN) Unconfigured Bad”
• “State change on PD … from ONLINE(18) to FAILED(11)”
• “VD 00/0 is now DEGRADED”
• “Foreign drives cannot be imported as they don’t form a complete configuration.”
• “Puncturing bad block on PD”
• MegaCLI commands that write to your drives
• Linux: “FW detected to be in faultstate, restarting it…”

If Your 9260-8i Array Is Down Right Now

1. If the volume is still accessible, copy the most important data off it before any rebuild, import or cooling work.
2. Photograph the POST and WebBIOS screens, or save MegaCli -CfgDsply -aALL, MegaCli -PDList -aALL and the event log to another machine.
3. Label every drive with its bay, which of the two connectors it hangs off, and any expander position; note serial numbers.
4. Do not import, clear, make good, force online, flash or rebuild — and do not move the drives to a PC.
5. Ship every drive together. We do not need the server or, usually, the card.

How We Recover 9260-8i Arrays

We image every member drive on write-blocked hardware in our ISO 5 Class 100 cleanroom, repairing physically damaged drives with donor parts first. Our HOMBRE software reads the MegaRAID metadata from each image, establishes member order, stripe size and parity layout, and works out which drives dropped out and when — which matters most on cards with a history of group drops. The array is assembled virtually, with no controller involved, and the file systems are recovered from the assembled volume. The full approach is described on our SAS2108 page.

9260-8i Documentation

• 6Gb/s MegaRAID SAS RAID Controllers User Guide (41450-04 Rev. D) — 9260-8i card layout, connectors, iBBU07 mounting
• MegaRAID SAS Software User Guide (Rev. I) — WebBIOS, MegaCLI, MSM and event messages
• MegaRAID SAS 9260CV Quick Installation Guide
• IBM ServeRAID M5015 and M5014 Product Guide

9260-8i Frequently Asked Questions

Several drives went “Unconfigured Bad” at once. Are they all dead?

Almost certainly not. Group drops on a 9260-8i point to a cable, connector, backplane, expander or power problem. If they all dropped at the same moment and are healthy, the documented make-good-and-import procedure usually works; if any failed earlier, image first.

My replacement 9260-8i will not import the old array. What now?

Disconnect the drives, check the replacement’s firmware, and match the original if possible. Do not clear the configuration to force it.

Can I use an IBM M5015 as a replacement?

It is the same chip, and arrays often move between them, but the firmware differs and RAID 6/60 need IBM’s feature key. On a degraded array, do not experiment.

Is it safe to run my 9260-8i without a battery?

Only in write-through mode. With write-back forced on and no working battery, a power loss discards cached writes.

My card runs hot. Is that a problem?

Yes. It needs active airflow. Heat shows up as resets and dropped drives before it shows up as a dead card.

What should I send you?

Every drive that was attached, labeled by bay and connector. We do not need the server, and in most cases not the card.

Start Your 9260-8i Recovery

Power the system down, label the drives, and do not import, clear, force, flash or rebuild. Our engineers will image every member and reconstruct the array without the controller. Gillware has been recovering RAID arrays for more than two decades.