Sun Fire X4500

The Sun Fire X4500 Servers (nick named "Thumper") are the heart of our $HOME. All user data reside on these boxes. To ensure that there are as few bottlenecks as possible, we distribute the users onto 10 of our X4500 and run daily snapshots to other boxes. If you are interested in the details please read on. This scheme allows us in the hopefully unlikely event of a whole box going down, to easily change configuration of a single file and give back the users' $HOME which should at most be 24hrs old.

Short Cuts

Serial numbers

General Description

• Where am I? On which server is my home directory?
• Backup strategy and other details...

HOWTO Jumpstart

Performance tests

• Performance evaluation for transferring ZFS snapshots

Problems (and solutions)

• mirroring system disks
• What to do, when Neterion cards don't use Jumbo Frame
• How to configure a replaced disk?
• Things to fix after a fresh Jumpstart install

ZFS

• Performance problems with NFSv3
• Possible zpool configurations

SUN cases

• Case ID38101108 - X4500 reboots infinitely when jumpstarting with mirror setup (status: reported, workaround given, but doen not work 100%)
• Case ID38104140 - X4500 does not detect Neterion XFrame II Fire card to enable PXE booting ( reported, Sun refuses to support, since this card is not Sun supported )
• Case 38102763 - ILOM access via SNMP and OIDs. The access is possible using the last firmware for the ILOM cards and the OIDs can be found in Clustermonitoring
• Case 38156975 - Getting fmadm fault information via snmp, the errors are visible via IPMI, technician will send more info by email
• Case 38118090 (via Marco Ehlert, MPI Colloid) - ZFS Patch for improved performance - needs new case
• Case 38185896 - multiple failure of slot in s08, what to do? Related Explorer is: /explorer-emea/explorer.27cf7a8b.s08-2008.10.10.12.14.tar.gz
• Case 70469164 - increasing ARP cache size
• Case 71257898 - after s11 hung error upon reboot, zpool import: cannot mount '/atlashome/BACKUP': directory is not empty
• Case 72065766 - zpool problems on s13, "random" vdevs are marked as degraded

-- CarstenAulbert - 12 Jun 2008