Dell CERC Adapter hangs temporarily until reset. REGRESSION
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
linux (Ubuntu) |
Triaged
|
High
|
Unassigned |
Bug Description
The HD CERC SATA 1.5 RAID controller (PCI-64 slot) hangs regularly for 10-60 seconds until reset.
The CERC card firmware is the newest I can find reference to: 4.1-0 [bld 7419]. The Dell Poweredge 1800 has newest BIOS (A07).
All hardware has been tested thoroughly OK except possibly the APIC component (Serverset Northbridge?).
The 14.04LTS with all updates (or not) works fine with a perfectly consistent throughput of 77MB/s (slowest HD=66MB/s), RAID5.
15.10 has the hangs of 10-60 seconds, AND with newest updates as of 26-FEB-2016.
I have changed PCI interrupts around and moved card slots to minimize sharing and removed cards - repeatedly.
All hard drives and cables have been thoroughly tested, but the drives are mixed: 1 of 66MB/s, 1 of 125 MB/s, and 3 of 100MB/s.
The drives not only work perfectly in 14.04LTS, but all Windows versions and DOS/BIOS-INT 19 as well (0x13).
I have tried all combinations of card RAID options Read ahead, Write cache, and other BIOS options.
Relevant dmesg lines (these were consecutive):
[ 1046.808027] aacraid: Host adapter abort request (4,0,0,0)
[ 1046.808123] aacraid: Host adapter reset request. SCSI hang ?
[ 1297.828038] aacraid: Host adapter abort request (4,0,0,0)
[ 1297.828168] aacraid: Host adapter reset request. SCSI hang ?
ProblemType: Bug
DistroRelease: Ubuntu 15.10
Package: linux-image-
ProcVersionSign
Uname: Linux 4.2.0-16-generic x86_64
ApportVersion: 2.19.1-0ubuntu3
Architecture: amd64
AudioDevicesInUse: Error: command ['fuser', '-v', '/dev/snd/seq', '/dev/snd/timer'] failed with exit code 1:
CasperVersion: 1.365
CurrentDesktop: Unity
Date: Thu Mar 3 05:08:57 2016
LiveMediaBuild: Ubuntu 15.10 "Wily Werewolf" - Release amd64 (20151021)
Lsusb:
Bus 001 Device 001: ID 1d6b:0002 Linux Foundation 2.0 root hub
Bus 004 Device 001: ID 1d6b:0001 Linux Foundation 1.1 root hub
Bus 003 Device 001: ID 1d6b:0001 Linux Foundation 1.1 root hub
Bus 002 Device 001: ID 1d6b:0001 Linux Foundation 1.1 root hub
MachineType: Dell Computer Corporation PowerEdge 1800
ProcEnviron:
TERM=xterm-
PATH=(custom, no user)
XDG_RUNTIME_
LANG=en_US.UTF-8
SHELL=/bin/bash
ProcFB: 0 nouveaufb
ProcKernelCmdLine: BOOT_IMAGE=
RelatedPackageV
linux-
linux-
linux-firmware 1.149
RfKill:
0: phy0: Wireless LAN
Soft blocked: no
Hard blocked: no
SourcePackage: linux
UdevLog: Error: [Errno 2] No such file or directory: '/var/log/udev'
UpgradeStatus: No upgrade log present (probably fresh install)
dmi.bios.date: 09/29/2006
dmi.bios.vendor: Dell Computer Corporation
dmi.bios.version: A07
dmi.board.name: 0P8611
dmi.board.vendor: Dell Computer Corporation
dmi.board.version: A04
dmi.chassis.type: 17
dmi.chassis.vendor: Dell Computer Corporation
dmi.modalias: dmi:bvnDellComp
dmi.product.name: PowerEdge 1800
dmi.sys.vendor: Dell Computer Corporation
Changed in linux (Ubuntu): | |
importance: | Undecided → High |
tags: |
added: regression-release removed: xeon |
Changed in linux (Ubuntu): | |
status: | Expired → Confirmed |
tags: |
added: kernel-bug-exists-upstream-4.7-rc4 removed: kernel-bug-exists-upstream-4.6 |
I forgot to add that although this is booted from the PXE at this time, I have previously installed and updated the 15.10.
I have tested both 32 and 64 bit versions with identical results, the server now has exactly 4G RAM (4x1GBxREGECC).
14.04LTS is currently installed as 2nd OS on HD, but it has a PCI NVIDIA card so the video doesn't work right (missing icons/menus/text in lists when moused over). I'm hoping the 16.04LTS will solve both issues simultaneously.
Both the 32 and 64 bit versions of 14.04LTS and earlier work OK.
The user is trying to use this server as a desktop workstation, so the onboard video isn't good enough, which is why it has the PCI card, it used to have 2 PCI video cards when he required DVI output because the BIOS won't recognize most cards, so no video until OS was booted. I have tried removing the PCI wireless also, which is the last remaining other card.
I do not specifically remember trying the drives with the onboard video only, but since it didn't matter if the interrupt was tied to the same line or not (tried in 2 different slots), I doubt that would make a difference.
Dan