The Server North
N

'blog

Home Main Site NOC/Info Site Twitter Stream Archives Login
You are currently viewing archive for January 2011
Jan
28
Strombo tells it like it is

http://openmedia.ca/meter

Posted by: Myke |

Jan
28
Crowbar's been pryed out Very early this morning we swapped out the hardware for Crowbar, from a Sun X2200 to an HP Proliant DL120 G6 - so far, so good. Perhaps this is the end of the KVM woes. We're holding tight while we try to prove a negative.

Posted by: Myke |

Jan
16
Fresh Blood We've ordered a new HP Proliant server to take over the Linux KVM based VPSes, since apparently the Opterons in the Sun Fire X2200M2s were too "old" for the task. (Who knew?)

Here's a mini-review of them (with photos!) on Myke's personal 'blog.

Hope to deploy after some burn-in, in the next week, along with the SAN upgrades.

Posted by: Myke |

Jan
15
Struggling with the SAN Finally figured out why the SAN crashed on New Year's Eve... and again today:
Double Panic - All The Way

The second OS drive failed, and since the HighPoint RocketRAID cards we're using are terrible pieces of... junk, they don't let us probe the drive's SMART status.

That will change in the next week or two, we're switching to dumb Marvell-based 8-port SATA cards that just come up as ATA or AHCI (depending) HBAs.

You might look at that 'screenshot' there (thanks ghoti!) and say "but that's just a file-system error, not a failed device" - well, maybe, but probably not. Either way, the HighPoints are out, then we can determine the drives' health directly.

Posted by: Myke |

Jan
04
Chainsaw's demise While we're not sure of the why the crash occurred, it's now clear why resetting the machine didn't work... the partition table on the boot-drive was corrupted.

Since we're big proponents of software RAID, I've got to admit this is where hardware RAID would've won. It seems that the entire bsdlabel was gone/corrupt on first drive (the second was fine), there was enough for the kernel to load, but not enough for root to mount and/or the mirror to fully load. (ie: use the second drive)

The second drive was fine however.

Either way, this was 2 too many crashes for this machine, the hardware is being retired. It was our oldest still-running machine - a Server North original in fact! Purchased almost exactly 6 years ago.

Newer hardware is carrying the name of Chainsaw.

Posted by: Myke |

Jan
02
Two Routers We're back to a redundant configuration.

Everything seems fine. ('cept for our OttIX connection, but that'll get fixed Real Soon Now.)

All Clear.

Posted by: Myke |

Powered by NucleusCMS | Ported by VinhBoy | Designed by DemusDesign