yumi got removed from the datacentre when I left Anchor, then was unhoused for quite a while. I was planning on selling off her hardware but she was in the boot when my car was written off and I was never able to collect her from the salvage company. RIP.

Specs

yumi is an IBM x3650 (model/variant: 7979B3M)

• 2-way Xeon E5420 @ 2.50GHz
• 6gb DDR2-667 FB-DIMMs
• IBM ServeRAID 8k adapter (not the 8k-l, this has the 256mb battery-backed write cache and more complex raid levels available)
  • 6x 750gb SATA - WD7500AACS
  • RAID-10 for 2tb of usable storage
  • Provides sda2 as an LVM physical volume

  • Use arcconf to interact with it

• 2x gigabit ethernet (Broadcom NetXtreme II BCM5708)

RSA

Remote management is through an RSA II Slimline card, it has it's own dedicated ethernet port.

You can talk to the RSA "in band", using the asu utility. Currently stashed in /root/ASU

• Dependent on the ibmasm service
• IBM are cocks, so of course there's only tarballs and RPMs for it
• That requires libusb and maybe something else

[root@yumi ASU]# ./asu64 rebootrsa

# if you forget the password:
[root@yumi ASU]# ./asu64 resetrsa

Default login is USERID // PASSW0RD (yes that's a zero in the password)

Certain crappiness is well documented: http://en.wikipedia.org/wiki/IBM_Remote_Supervisor_Adapter#Peculiarities

Forgotten password something-something:

• http://brain4free.org/wiki/doku.php/blog:reset_password_on_a_ibm_rsa_ii

• http://webcache.googleusercontent.com/search?q=cache:wOY_UfBo4hsJ:www.ibm.com/developerworks/forums/thread.jspa%3FthreadID%3D165997+&cd=2&hl=en&ct=clnk

Here's the manual: asu_guide.pdf

Network config

./asu64 set RSA_Network1 Enabled
./asu64 set RSA_HostIPAddress1 202.4.232.66
./asu64 set RSA_HostIPSubnet1 255.255.255.0
./asu64 set RSA_GatewayIPAddress1
./asu64 set RSA_GatewayIPAddress1 202.4.232.254
./asu64 set RSA_DHCP1 Disabled
./asu64 rebootrsa

Drive problems

As of March 2014, it was identified that some drives are crapping out. Between reboots, 3 of the drives can be expected to disappear and reappear for no good reason. Clearly the drives are flaking out or something (or the backplane/connectivity is bad).

There are six bays, numbered 0 to 5. Initially the middle pair was apparently toast (killed the RAID-1 pair in the middle of the RAID-10 set), but it looks like the last drive isn't good either. That is, bays 2/3/5 look bad.

As a test, I swapped the top row (0/1) with the middle row (2/3), and bay 4 with bay 5. This should give me a better idea of whether it's a drive problem vs. a connection problem.

On the current boot (2014-03-14, 16:20), three drives are present and functional. The serial numbers according to arcconf getconfig 1 are:

• WD-WCASM0050857 (slot 2)
• WD-WCASM0039771 (slot 3)
• WD-WCASM0039771 (slot 5)

Given that I swapped 3 good drives with 3 bad drives in those positions, this suggests that the drives are the problem.

IPv6

yumi is 2407:7800:0100:01c9:0000:0014:b022:f5dd/64

jails are in 2407:7800:0100:01c9:0000:0014:b022:f800/117 (f800 to ffff in the last quad)

jails also include 202.4.233.48-55 (in /24 subnet)