boot halts when SATA port multiplier removed or no disk attached to port 0
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
TurnKey Linux |
Won't Fix
|
Low
|
Unassigned | ||
11.x |
Confirmed
|
Low
|
Unassigned |
Bug Description
When using TurnKey Linux version 11.1 (= Ubuntu 10.04.1 LTS X86) + booting from internal drive + external storage on Highpoint RR2300 PCIe X1 SATA controller + 2 x Dawi / Dawicontrol DC-6510 PM SATA port multipliers + all connected disks are used in non-RAID alias Legacy disk mode + DHCP networking, the Linux boot process halts when (1) the number of port multipliers connected is reduced from 2 to 1 (2) there is no longer a disk connected on port 0 of any connected port-multiplier.
Problem:
The boot process halting makes any remote tools unavailable: no webmin && no ssh.
Expected result:
An uninterrupted boot process on any SATA hardware changes. As hard drives can fail anytime that should not block the boot process including remote admin tools.
Regression:
Switching from DHCP assigned to fixed/static IP networking, does not improve.
Workaround:
(1) (Hot-)plug an identical number of port-multipliers as on OS installation moment.
(2) (Hot-)plug any SATA disk device on port 0 of any connected port multiplier.
Notes:
The boot process on the attached monitor displays: "udhcpc[798]: Lease of 192.168.254.8 obtained, lease time 86400" as last line.
After connecting port-multiplier
the next outputted lines are: "udhcpc[1007] udhcpc (v0.9.9-pre) started" and "udhcpc[1007]: SIOCGIFINDEX failed!: No such device" and " * Starting Initialization hooks"
1 second later the screen is cleared, and displays "[ 6.120510] Console: switching to colour frame buffer device 160x64",
4 seconds later the screen is cleared again,
3 seconds later the blue "TurnKey Linux Configuration Console" is shown.
Logs (kern.log, dmesg) from 19:14GMT+01 started but halting, wait for half a minute, disk connect on 2nd multiplier port 0, boot process are attached.
Changed in turnkeylinux: | |
status: | New → Triaged |
status: | Triaged → Confirmed |
importance: | Undecided → Low |
I have changed this to "Won't fix" as it applies to v11.x. Now that v12.0 has been rebased on Debian, hopefully it is no longer an issue. If someone finds that it is still an issue, then please target v12.x and change this bug back to "New".
Thanks