Hardware Migration [11/02/2009] - [11/05/2009]


Recommended Posts

We will be migrating the Jedi node to new hardware starting on November 1st and it will last up until November 5th although we estimate it'll take not even 24 hours to migrate the entire node to new hardware.

There will be no down time during this migration we will move VPS's using the migration tools in OpenVZ to minimize the impact of the move as much as possible. We're sorry about the short notice regarding this we have been monitoring the health of this node and we feel it should be upgraded as soon as possible to restore it to it's optimal performance it once had.

Specification break down:

Old Jedi

Dual Xeon 5430


4x300GB SATA Raptor 10K RPM Raid-10

New Jedi

Dual Xeon 5520


4x300GB SAS 15K RPM Raid-10

So this is a significant jump in specifications on this node. We have switched to using Xeon 5520's which are much faster than a 5430 CPU. We are also doubling the amount of ram on the system so there is further caching is involved. We are also using the much faster and more reliable serial attached scsi drives over the SATA raptor drives we were using. The raptors are faster than most providers use for drives but we think at this point we should be using SAS drives for all our servers.

If you have any questions or concerns about this migration do not hesitate to make a VPS ticket.

Start Date: 11/02/2009

End Date: 11/05/2009

Estimated Down Time: less than 5 minutes

Duration: 4 days

Link to comment
Share on other sites

We were about 60% completed when we had to stop due to finding something odd with the new server. The machine passed our burn in tests but the logs suggest there is a bad stick of memory.

Rather than moving everyone back then fixing the memory we're just going to take the 20-30 minutes and replace the bad stick. The reason being the time it takes to move everyone back and having to do it when memory is bad is not a good idea. We'll be doing this ASAP and not all users are affected either. Unfortunately no way around this the memory is throwing an increasing number of errors.

Link to comment
Share on other sites

This topic is now closed to further replies.