[dm-devel] - info about dm

Vega Forneris vega.forneris at sic.rm.cnr.it
Wed Nov 7 13:51:52 UTC 2007


Dear all,

I just installed Scientific Linux 5 and I used for the first time DM
drivers for working with the embedded RAID (the servers are IBM xSeries,
with an Adaptec HostRAID ASR - SATA). I really appreciate the fact that
now Linux is able to recognize such devices, but I had some troubles
configuring it and I’d like to have some info for better understanding
how things work. 

I’ll be a little “verbose” for clarifying what I did and which are my
doubts


1) Installation & Configuration
I installed SL5 on a system which was running SL3 with HW drivers
released by IBM. Linux was able to identify correctly the device
(installed through CD), mapped it as /dev/mapper/asr_ and proceeded with
he installation.

Anyway, I found a little problematic configuring it manually: I tried to
install SL5 on a node which was misconfigured: instead of seeing one
device, it was managing the 2 HD (/dev/sda and /dev/sdb) as 2 separate
devices an if I tried to configure them (at install time, via CD or
network boot) with  DM, I got many errors about the RAID table (“ERROR:
asr: Invalid RAID config table checksum on /dev/sda” or even “asr:
Invalid magic number in RAID table; saw 0x0 expected
”). I tried to
reconfigure the RAID card first via BIOS and then trying again to
configure it via Linux (“dmsetup create 
”) but even if it seemed
working (exit code 0), then querying the device asr_ (i.e. “dmraid –r”)
resulted in errors as reported above
or better, dmraid commands (-r,
-ay, -tay,
) reported errors, while dmsetup commands  (status, ls,
)
worked normally. The only way I found for well-configuring the device
was to install SL3 with specific drivers, and then to install SL5 with
DM
both commands now work perfectly

In brief:

If I try to configure DM-RAID from scratch and I get such errors, how
can I do? What’s that “RAID table”? Where is it written? Not in the MBR
of an HD, ‘cause I tried to copy the MBR of the cloned disk which had no
problem (still recognized as “good” for RAID and DM) and nothing
changed


2) Managing disks
I configured disks in RAID 1 (mirror) but I wonder how to manage this
RAID; in particular, how can I do for:
 - stopping a device (disk) and put it offline/and viceversa? 
 - reconfiguring/rebuilding it (due to broken/faulting HD)? 

At the moment I should change a couple of disks (SMARTD warnings
reported), but I’m scared about the “RAID table error”


Many Thanks in advance for all help you can give me


Best Regards
Vega Forneris

+--------------------------------------+
 Vega Forneris

 CNR – Servizi Informatici di Campo
 Gestione servizi GRID

 Area di Ricerca Roma II - Tor Vergata
 Via del Fosso del Cavaliere, 100
 00133 Roma

 Tel:   +39 06 4993 4667
 Fax:   +39 06 4993 4650
 Email: Vega.Forneris at sic.rm.cnr.it
+--------------------------------------+






More information about the dm-devel mailing list