That server’s a monster… if it wasn’t for the occasional power outage (the ‘blessing’ of lake-effect-snow storms) and hardware failures, these uptimes would be better. Regardless, they’re not bad in the least. The list of “1 year uninterrupted uptime” for Unix / Linux servers far outshines that of Win-based units — even though Windows systems eat up the majority of the marketplace / installed systems.
Here are two of our own… pentiumslayer1.spinellicreations.com and storageslayer.spinellicreations.com – both running RHEL 5.4 derivatives (Fermi CERN’s Scientific Linux).
These aren’t slouches, but they’re a far cry from SUN SPARC’s. For example, we’ll have to reboot in the next few weeks or so, after a failed RAID-disk has been replaced (both of them have lost a disk over the past few months – the benefit of RAID-6, no worries). While the drives and backplane are hot-swappable, the host (disk) adapter is not. Perhaps the point that should be made here is that, were it not for the occasional hardware failure needing service, these systems would basically never be rebooted — a testament to the versatility (in terms of ability to upgrade / make changes on the fly without rebooting the OS) and robust performance (rock solid kernel stability results in an un-degraded system over time) of the Linux OS.