ProLiant DL360p Gen8 + P421 + D2700 enclosure + hpsa - not seeing logical disk device (Ubuntu 14.04)

Question

I have inherited a ProLiant DL360p Gen8 with two Smart Array controllers:

# lspci -d '103c:323b' -k
02:00.0 RAID bus controller: Hewlett-Packard Company Smart Array Gen8 Controllers (rev 01)
        Subsystem: Hewlett-Packard Company P420i
        Kernel driver in use: hpsa
04:00.0 RAID bus controller: Hewlett-Packard Company Smart Array Gen8 Controllers (rev 01)
        Subsystem: Hewlett-Packard Company P421
        Kernel driver in use: hpsa

The P421 controller apparently has a D2700 enclosure plugged into with a bunch of SAS drives configured as RAID6:

# hpssacli ctrl slot=1 show config                                                                                                                                                      
Smart Array P421 in Slot 1

   StorageWorks D2700 at Port 2E, Box 1, OK
   array A (SAS, Unused Space: 0  MB)

      logicaldrive 1 (19.1 TB, RAID 6, OK)

      physicaldrive 2E:1:1 (port 2E:box 1:bay 1, SAS, 1 TB, OK)
      physicaldrive 2E:1:2 (port 2E:box 1:bay 2, SAS, 1 TB, OK)
      physicaldrive 2E:1:3 (port 2E:box 1:bay 3, SAS, 1 TB, OK)
      physicaldrive 2E:1:4 (port 2E:box 1:bay 4, SAS, 1 TB, OK)
      physicaldrive 2E:1:5 (port 2E:box 1:bay 5, SAS, 1 TB, OK)
      physicaldrive 2E:1:6 (port 2E:box 1:bay 6, SAS, 1 TB, OK)
      physicaldrive 2E:1:7 (port 2E:box 1:bay 7, SAS, 1 TB, OK)
      physicaldrive 2E:1:8 (port 2E:box 1:bay 8, SAS, 1 TB, OK)
      physicaldrive 2E:1:9 (port 2E:box 1:bay 9, SAS, 1 TB, OK)
      physicaldrive 2E:1:10 (port 2E:box 1:bay 10, SAS, 1 TB, OK)
      physicaldrive 2E:1:11 (port 2E:box 1:bay 11, SAS, 1 TB, OK)
      physicaldrive 2E:1:12 (port 2E:box 1:bay 12, SAS, 1 TB, OK)
      physicaldrive 2E:1:13 (port 2E:box 1:bay 13, SAS, 1 TB, OK)
      physicaldrive 2E:1:14 (port 2E:box 1:bay 14, SAS, 1 TB, OK)
      physicaldrive 2E:1:15 (port 2E:box 1:bay 15, SAS, 1 TB, OK)
      physicaldrive 2E:1:16 (port 2E:box 1:bay 16, SAS, 1 TB, OK)
      physicaldrive 2E:1:17 (port 2E:box 1:bay 17, SAS, 1 TB, OK)
      physicaldrive 2E:1:18 (port 2E:box 1:bay 18, SAS, 1 TB, OK)
      physicaldrive 2E:1:19 (port 2E:box 1:bay 19, SAS, 1 TB, OK)
      physicaldrive 2E:1:20 (port 2E:box 1:bay 20, SAS, 1 TB, OK)
      physicaldrive 2E:1:21 (port 2E:box 1:bay 21, SAS, 1 TB, OK)
      physicaldrive 2E:1:22 (port 2E:box 1:bay 22, SAS, 1 TB, OK)
      physicaldrive 2E:1:23 (port 2E:box 1:bay 23, SAS, 1 TB, OK)
      physicaldrive 2E:1:24 (port 2E:box 1:bay 24, SAS, 1 TB, OK, spare)
      physicaldrive 2E:1:25 (port 2E:box 1:bay 25, SAS, 1 TB, OK, spare)
[...]

However, the configured logicaldrive doesn't show up in the SCSI subsystem:

# lsscsi -g
[2:0:0:0]    disk    HP       LOGICAL VOLUME   6.34  /dev/sda   /dev/sg1 
[2:3:0:0]    storage HP       P420i            6.34  -          /dev/sg0 
[3:0:0:0]    disk    HP       LOGICAL VOLUME   6.34  -          /dev/sg3 
[3:3:0:0]    storage HP       P421             6.34  -          /dev/sg2

The logicaldrive was initially in a failed state which I re-enabled via:

# hpssacli ctrl slot=1 ld 1 modify reenable

after which I figured the drive would show up in the system, which it didn't (no new messages in dmesg either).

I also did a re-scan:

# echo 1 >/sys/class/scsi_host/host2/rescan
# echo 1 >/sys/class/scsi_host/host3/rescan

with no effects.

Am I missing a fundamental step (I have not rebooted the server/enclosure...yet)?

# fdisk -l
Disk /dev/sda: 100.0 GB, 99994337280 bytes
255 heads, 63 sectors/track, 12156 cylinders, total 195301440 sectors
Units = sectors of 1 * 512 = 512 bytes
Sector size (logical/physical): 512 bytes / 512 bytes
I/O size (minimum/optimal): 262144 bytes / 262144 bytes
Disk identifier: 0x0005a4c4

   Device Boot      Start         End      Blocks   Id  System
/dev/sda1   *         512      500223      249856   83  Linux
/dev/sda2          500734   195300863    97400065    5  Extended
Partition 2 does not start on physical sector boundary.
/dev/sda5          500736   195300863    97400064   8e  Linux LVM

Disk /dev/mapper/sys1--vg-root: 31.0 GB, 31016878080 bytes
255 heads, 63 sectors/track, 3770 cylinders, total 60579840 sectors
Units = sectors of 1 * 512 = 512 bytes
Sector size (logical/physical): 512 bytes / 512 bytes
I/O size (minimum/optimal): 262144 bytes / 262144 bytes
Disk identifier: 0x00000000

Disk /dev/mapper/sys1--vg-swap_1: 68.7 GB, 68681728000 bytes
255 heads, 63 sectors/track, 8350 cylinders, total 134144000 sectors
Units = sectors of 1 * 512 = 512 bytes
Sector size (logical/physical): 512 bytes / 512 bytes
I/O size (minimum/optimal): 262144 bytes / 262144 bytes
Disk identifier: 0x00000000

Output of hpssacli ctrl all show config defailt here.

Show the output of `fdisk -l` - Also http://pastebin.com the output of `hpssacli ctrl all show config detail`. Thank you. — ewwhite, Mar 22 '16 at 12:41

ewwhite · Accepted Answer · 2016-03-22T13:24:48.303

2

What was initially "failed" about the array? This is one of those cases where I'd like to know what happened just before you wrote the question.

[3:0:0:0]    disk    HP       LOGICAL VOLUME   6.34  -          /dev/sg3

Your D2700's logical drive is there. Maybe a dmesg | tail will show you the rescan information following your modify reenable command. I can definitely say that the reenable switch is not commonly used. It's not something that needs to be used in normal operation. So:

Why the the array fail? Was the D2700 physically unplugged?
Where are you expecting this to be mounted? Look in /etc/fstab to see if there's a mountpoint waiting for this logical drive.
Reboot (or help us understand why you haven't rebooted yet).

edited Mar 22 '16 at 13:24

answered Mar 22 '16 at 13:18

ewwhite

194,921
91
434
799

Very good questions indeed, sadly I don't know the answers. The logicaldrive had already failed when I first accessed the server with all array member drives being OK according to `hpssacli` so I can only speculate what happened (I read [this question](http://serverfault.com/q/352502/167825) and your answer before I re-enabled the drive). There was *no* output in `dmesg` after running `modify reenable` or any other command whatsoever. – Adrian Frühwirth Mar 22 '16 at 13:28
I did get bogus output after running `echo "- - -" >/sys/class/scsi_host/host[23]/scan` but from what I understand that's a non-issue (https://bugzilla.redhat.com/show_bug.cgi?id=1139407); nothing after `echo 1 >/sys/class/scsi_host/host[23]/rescan`. Checking `/etc/fstab` was the absolute first thing I did but from what I understand the enclosure has not yet been in use since the logical drive never showed up. I haven't rebooted the server yet because downtime might be an issue and I assumed it *should* not be needed, but I'll gladly check back and do so if that might resolve the issue. – Adrian Frühwirth Mar 22 '16 at 13:30
1

Delete the logical drive and recreate it if it doesn't have any data on it. – ewwhite Mar 22 '16 at 13:49
I hate to say it but I had to recreate the logical drive and all is working now. Thanks! I do wonder what had happened to the array, though ... – Adrian Frühwirth Mar 22 '16 at 19:59

ProLiant DL360p Gen8 + P421 + D2700 enclosure + hpsa - not seeing logical disk device (Ubuntu 14.04)

1 Answers1