14 thoughts on “Unscheduled Maintenance Friday, January 25, 2013 2-3pm PST

  1. It is when they have downtime to fix the last fix which was to fix the previous mistakes in the fix before that which makes people angry. Downtime is expected, but multiple down times to fix things that they screwed up the first time is annoying.

    On another note, GW2 has a nice system where the server patches while its running and just prompts players to reboot the game when it is done. No downtime, no waiting other than the time it takes you to download the patch.

  2. Aye Downtime is to be expected but not excessive downtime due to lack of ability to make things work the first time. SOE please get some better coders and make things work the first time upon release. Then I can stop bitching like seriously said, you know since I guess I should have dinner. WTFE

  3. On another note, GW2 has a nice system where the server patches while its running and just prompts players to reboot the game when it is done. No downtime, no waiting other than the time it takes you to download the patch.

    I’m not here to extol the virtues of GW2 over EQ2 (which I’ve now quit) but this certainly is a welcome way of doing things on new-build day. I don’t have enough knowledge of the technology but perhaps it’s something SOE could think about employing.

  4. Patching the client without taking the servers down is one thing, but patching servers is an altogether different beast.

    In this case it is not just ‘a server’, as in singular; and probably not just ‘a game patch’. They probably have to do other things, like patch the server os, and any related third-party services/activities. Then to keep things sane, they probably do some verification to ensure the patches got applied and everything is working mostly harmoniously. Until brought back every potential problem with the patch can’t be identified, but I am sure they try to keep problems to a minimum, things that are taken for granted. So aside from copying and overwriting files, they probably do at least a moderate amount of testing even when everything seemed to work just fine in a testing/qa environment.

    Another part of this equation that may not be readily obvious is any changes in data structures or underlying storage and messaging/communication formats for game and player data. Again, techniques and technology exist to enable on-the-fly updates to format, but the tradeoffs in complexity and performance might really diminish the value of on-the-fly data migrations.

    There certainly are techniques and technologies to make seamless, on-the-fly runtime/server ‘code’ updates possible, but like everything else, they come with tradeoffs, and further, it probably wasn’t a key objective when they built the system. This kind of change would fundamentally change their software design/architecture… something that would be unlikely at this stage of the game for them and us, when other issues tend to be more pressing and more valued in the long run, its probably more productive to thank them for their dedication and progress in improving what we have — even if we are paying them for a service.

    I am sure the qa/test tool for test/beta that applied changes to a live environment while immensely helpful to ‘live-tuning an encounter’ is vastly less complicated, especially in light of the fact that any changes it made were temporary till the zone or server came down. The use of the word ‘patch’ was somewhat disengenous, but at the same time, for all intents and purposes, it was a patch, just not very ‘sticky’. It was more like a couple of safety pins that would hook and hold key points together till the item could be properly mended. What surprised me the most from the announcemnt of that tool is that they didnt already have that kind of capability from day one.

Leave a Reply