Hotswap discs Handling

[Daniel Bareiro]

Hi all!

I'm using Ubuntu server LTS 8.04 with HP Proliants equipment. In particular, they are DL380 G5 with four SAS discs in RAID 5.

My doubt aims at if it is necessary to execute some previous command at the time of remove one of discs and after to put some new one. Searching in Internet I found this [1] document, although I didn't find the archives of Linux kernel that are mentioned there.

I could obtain that hpacucli works on Ubuntu GNU/Linux and it interacts with the Nagios plugin check_cciss and now I wanted to evaluate the degree of information from them, removing one disk and returning back to the server ('degraded' and 'reconstructing' states would have to be shown, I suppose).

But it wanted to know if it is necessary to execute some command from hpacucli or on /proc so that Linux kernel does 'rescan' of the discs, although I do not have independent physical devices here but the device /dev/cciss/c0d0p*

In an equipment Sun Enterprise 220R in which Debian GNU/Linux Sparc are used, to remove and to add SCSI discs I used the utilities of the package 'scsiadd' and for that reason I thought that under these circumstances it would be necessary to make something similar.

Thanks in advance.

Regards,
Daniel

[1] http://www.dufault.info/blog/hotswap-a-scsi-sas-or-sata-drive-in-linux/

[David Claypool · Accepted solution]

It's totally transparent to the kernel. The SmartArray controller is hardware RAID and nothing is necessary to inform the OS about any changes. There actually is a sensor in the drive bay that tells the controller of the change.

[Daniel Bareiro]

Thanks for your reply, David. Then now I'm safe that I can remove the disc without to perform some previous operation on the OS.

The next week I will review the information degree that I obtain with check_cciss Nagios plugin and hpacucli after to remove a disc and when returning it to place.

At the moment I obtain the following output with check_cciss Nagios plugin:

# /usr/local/nagios/libexec/non-std/check_cciss -v
RAID OK: Smart Array E200 in Slot 3 array A logicaldrive 1 (410.1 GB, RAID 5, OK) [Controller Status: OK Cache Status: OK Battery Status: OK]

And trying to obtain a little more data, I execute hpacucli with the following syntax:

# hpacucli controller slot=3 physicaldrive all show

Smart Array E200 in Slot 3

array A

physicaldrive 1I:1:1 (port 1I:box 1:bay 1, SAS, 146 GB, OK)
physicaldrive 1I:1:2 (port 1I:box 1:bay 2, SAS, 146 GB, OK)
physicaldrive 1I:1:3 (port 1I:box 1:bay 3, SAS, 146 GB, OK)
physicaldrive 1I:1:4 (port 1I:box 1:bay 4, SAS, 146 GB, OK)

Two doubts that arose to me on the tests that I will do:

* If for example I remove the disc on bay 3 trying to simulate a failure in one of discs of RAID 5, the execution of the previous command would have to show that the disc on bay 3 is inactive?

* After to return to put the disc in bay 3, I suppose that I would have to execute some operation with hpacucli to add the disc to RAID 5 and rebuild the array. Is it correct?

An isolated question: of the output of check_cciss Nagios plugin I see ' Array A' and ' logicaldrive 1'. They are equivalent nomenclatures?