RainDrop Laboratories Status

Peak replaced its core routers last night, and in the process, the route to my network got lost due to an obscure configuration issue.  It should be fixed shortly…

There were some issues getting the equipment into the rack, but managed a work around, copied vm files and up it all came.  Only problem is an aggravated time issue with the mail server; I’m building a new version with a more forgiving setting so it’ll be happy again…  All services should be operational now…

I’ve postponed the vmserver upgrade a week in order to get a couple more pieces of hardware for it.  I’ll be working on the new services on it instead, so no disruptions this weekend…

Saturday afternoon, I’ll be migrating the virtual machines to a new host machine.  This requires format conversion, which takes about 1/2hr-1hr for each one (but each one will be up until its conversion process begins, and will then come back up immediately).  The main impact will be a nameserver and outgoing mail.  Once this is done, I will be able to move additional services off the old agora and onto new vms.

For the technically inclined, I’m upgrading from a 2.8Ghz P4/Xeon w/3G ram running vmware server to a 2.83Ghz Q9550 quad-core w/8G ram running ESXi.  The new vms will reside on an opensolaris zfs/nfs server as well.  The only storage on the vmhost is a 32G ssd.

There was another problem with the nameserver, relating to earlier work on supporting secure dns.  This caused the problem with emu, and probably the performance problems some people have been seeing.  Note that while emu is now working, it will be retired when I get the new mail server built…

Emumail seems to have lost its mind in the move; I’m not sure how, as nothing inside the network should be able to tell that it didn’t just reboot.  In any case, it’s been on its last legs for some time, and this is a good opportunity to retire it for good, and good incentive for me to get the new mail server going…

All went about as smoothly as could be expected.  Agora is a little finicky about its power up sequence these days, and the virtual machines noticed things had changed (not sure what though, they shouldn’t have seen any change at all!), and so needed some comforting to start up, but all are happy now.  I will be finalizing a new virtual machine host over the next week, and plan to move the vms to it next weekend.  Once that’s done, I’ll probably move the web server from agora to its own vm next, followed by a new shell server and finally mail, which should allow the retirement of the antique hardware finally!

There will be a brief network outage this afternoon while I upgrade
the router firmware in preparation for supporting IPv6 (a new Internet
protocol).  There *may* be a more extended outage for agora itself,
depending on if a tool to copy it into a virtual machine works or not.
This may be a way of getting it off the questionable hardware it’s
running on currently…if it works…

Another bullet dodged — it just rebooted…

I’ve got the colo manager checking it out, but it is likely the complete hardware failure it’s been threatening to do, which means it will be late tonight at the earliest before it’s back up…