Description of problem: I have a system with an Overland Powerloader AIT3 tape library hanging off a Tekram DC-390U3W controller. With recent kernels, this system becomes unbootable. Trying to boot kernel-smp-2.4.21-27.0.2.EL results in a string of SCSI bus resets and finally an OOPS. Trying to boot kernel-smp-2.4.21-27.0.1.EL results in (seemingly) neverending SCSI bus resets. kernel-smp-2.4.21-20.0.1.EL works just fine. I'm using the sym53c8xx_2 driver, which seems to be the same version among these kernels. System info: dual 2.4 GHz Xeons on Supermicro X5DPI-G2 motherboard. lspci info for Tekram controller: 02:01.0 SCSI storage controller: LSI Logic / Symbios Logic 53c1010 Ultra3 SCSI Adapter (rev 01) Subsystem: Tekram Technology Co.,Ltd. DC-390U3W Flags: bus master, medium devsel, latency 140, IRQ 48 I/O ports at 3000 [size=256] Memory at fc204000 (64-bit, non-prefetchable) [size=1K] Memory at fc200000 (64-bit, non-prefetchable) [size=8K] Expansion ROM at <unassigned> [disabled] [size=64K] Capabilities: [40] Power Management version 2 02:01.1 SCSI storage controller: LSI Logic / Symbios Logic 53c1010 Ultra3 SCSI Adapter (rev 01) Subsystem: Tekram Technology Co.,Ltd. DC-390U3W Flags: bus master, medium devsel, latency 140, IRQ 48 I/O ports at 3400 [size=256] Memory at fc204400 (64-bit, non-prefetchable) [size=1K] Memory at fc202000 (64-bit, non-prefetchable) [size=8K] Expansion ROM at <unassigned> [disabled] [size=64K] Capabilities: [40] Power Management version 2 cat /proc/scsi/scsi: Attached devices: Host: scsi0 Channel: 00 Id: 03 Lun: 00 Vendor: OVERLAND Model: LIBRARYPRO Rev: 0420 Type: Medium Changer ANSI SCSI revision: 02 Host: scsi0 Channel: 00 Id: 04 Lun: 00 Vendor: SONY Model: SDX-700C Rev: 0202 Type: Sequential-Access ANSI SCSI revision: 02 Host: scsi0 Channel: 00 Id: 05 Lun: 00 Vendor: SONY Model: SDX-700C Rev: 0202 Type: Sequential-Access ANSI SCSI revision: 02 Version-Release number of selected component (if applicable): How reproducible: Always Steps to Reproduce: 1. Boot system with recent kernel Actual results: SCSI bus resets/OOPS Expected results: Normal operation Additional info:
The version of the RHEL3 kernel which was last released was kernel-smp-2.4.21-27.EL, not the versions you mention above. Please check if you can repeat the problem on that kernel.
Actually, 2.4.21-27.0.2.EL is the latest release (2nd security errata release after U4), so this is fine.
Joshua, You are right, there were no changes to the driver. There were also no changes to the SCSI midlayer, including the SCSI whitelist, that look relevant. Are the SCSI bus resets preceeded by command timeout messages? Please post the messages. My guess is that interrups are not getting delivered, causing command timeouts, which cause bus resets. Let us know what type of system it is. Tom
Well color me confused. Now that I'm trying to capture logs, I can't recreate the bug. 27.0.1 and 27.0.2 boot in the same manner 20.0.1 did. What's even better is that I haven't touched the library since those kernels wouldn't boot. Before I opened this bug, I had tried all the obvious things (checked cables, power-cycled the loader, etc), and it still wouldn't boot. Now, a few days (and backup cycles, but no power cycles of the loader) later, it seems to work just fine. I guess I'll close this and mark it as a Heisenbug. Sorry for the noise.