Monday's (Oct 3 2011) restarts

[DavidThomas Scorbal]

"..will require an extended period of region restarts. We expect this maintenance to last at least a week."

I'm not clear on the meaning of this ominous sounding warning. 

Does this mean that each region will be restarted several times throughout the week? Or that each region will be down for longer than normal, requiring a longer overal restart process from start to finish?

[Ayesha Askham]

@DavidThomas

I suspect you may find that both eventualities will apply.This sounds like no small undertaking and is probably long overdue.

Recent LL software rolls have not gone well, but I think that this period, being so critical, will be very closely watched.  I also fear (though I admit I have no grounds in evidence for this) that some viewer 1 functions may be impaired by these upgrades.

After all, V2 and its spawn have been with us for a long time now, and we have been given numerous warnings that optimisation of SL for V2 would degrade the usability of V1s.  I should at this stage admit to being a recidivist V1 user!

I think we are in for a bumpy ride this coming week, and we have no choice but to sit it out and wait and see!

 

PS: I find it a bit annoying, though par for Linden Lab's course, that there has been no article in the Official Blog about this, considering that it will, by its nature, affect all of SL.

[Void Singer]

I expect it just means "expect to see extra restarts during the week" as they take one machine to update it, the region gets restarted elswhere.... once you end up on one of the upgraded machines your region should be safe from those, but I imagine that will take several shuffles for some regions (it'd be nice if they could target them so that blank or empty linden regions are the first to go, but they're probably cross hosted with heavy use regions)

[WolfBaginski Bearsfoot]

It does seem a little carelessly worded, but if it is an upgrade of the OS software it will take longer, because there's more data to move. I'd hope that some server hardware is already running the new OS version so that they can use a standard disk image that will apply to all the servers.

This should be a very ordinary sysadmin job.

[Lydia Craig]

Three things would be extremely helpful.  First, LL needs to go back to the old restart warning system, stip relying on open chat which is often missed if you are buy in IMs, building or doing virtually anything but scanning chat.  Two it would be nice if they would tell us what they intend to accomplish by this move, what we should expect while its it going on, and when servers are down for unusual times give us reports on their current condition.  Finally, it would be extermely good if LL provided a post event team to help with any problems that may arise during this process.

Sadly, this is expecting a bit much of LL at the moment and we likely will be met with the same attitude as usual and given empty or vague information and meaningless promises, and the result will be another exodus of players and customers from SL.

[Void Singer]

---

Lydia Craig wrote:

[...] Two it would be nice if they would tell us what they intend to accomplish by this move, what we should expect while its it going on, and when servers are down for unusual times give us reports on their current condition.  Finally, it would be extermely good if LL provided a post event team to help with any problems that may arise during this process.[...]

---

they did tell you... you just didn't understand.... OS upgrades... which is just updating the machines that SL runs on, not SL itself (security patches, maybe some minor improvements on how they work), and what's going to happen (extra region starts, caused by them swapping out machines that those regions are running on).

since restarts are all that should be happening, regions shouldn't be down for excessive amounts of time. if they are, treat it the same way you treat any restart that doesn't come back up... report it to live chat, or via a ticket (on of the few ticket types that gets almost immediate response). and if something goes horribly wrong they'll post it in the gridstatus page as usual (not that I think that's effective or sufficient for a great many things, but it's what's likely to happen)

 

I do prefer the popup notices for restarts myself too though.

[Nalates Urriah]

As to the Lindens telling us what they hope to accomplish... they do. Most residents are not interested. If the Lindens put all the information they have out to the whole population, people would stop paying attention, its TMI. We would be in the same situation as if they did not tell us. Information overload...

The way it is handled is most of the less technical and more general interest news hits the SL Forum and Blog. Some of the more tech stuff of general interest is announced in the tech sections of the forum.

The detailed tech stuff is announced and/or discussed in the User Group meetings.

The extra restarts are required because they are upgrading the operating system kernel on the simulator and support servers Oct 3 to 7. This is an effort to fix the TimeWarp problem.

The regular main channel Tuesday restart will likely happen as a release channel update passes testing and those fixes get rolled to the main channel/grid. If all 3 RC's fail testing, that restart will be avoided. The release channels will see the regular Wednesday restarts as they get new or revised updates to test.

We still get the 5 minute warning in-world prior to a region closing down with follow on warning for late arrivals. It is fun to watch people at large gatherings run to an adjacent region when a warning pops. They break out the popcorn to watch the new residents scratching their heads wondering where everyone went then poof when the region restarts. 

For a summary of the technical side of what is going on in SL follow my blog.

 

[Ayesha Askham]

@Nalates

The real gripe here is that NO Official info, other than on the Grid Status page has been given at all.  Yes those interested can probably drill down far enough to find it, but otherwise it is down to ordinary residents to pass out info...and THAT is how rumours start.

@Void, the pop-up was so much more noticeable than the chat line, you are right, but LL cannot win there I recall folk complaining of "How irritating" it was <sigh>.

Overall I think that while LL is improving its communication with its SL users, it still has some way to go.

I do not agree that some minimal info on the Official Blog is TMI, it is far better than unfounded rumours.

[Deej Kasshiki]

So it's better to say nothing and expect those affected and interested to hunt down the info scattered over several LL web properties or THIRD PARTY sites? Attend User Group meetings held during times when most people are at work?

 I can't believe you're actually defending the Lab's opaque and ineffective comunication style.

I'll grant you that most residents have no understanding or concern for the more technical aspects but, it wouldn't be too much to expect some general, easy to understand information about these issues that affect the entire grid to be published in a more centralized (and VISIBLE)  location on an official Linden property.

 

 

[Josh Susanto]

It means that, for some substantial period of time, your paid listing enhancements will not result in additional sales.

Of course, they couldn't have told you this BEFORE you bought those enhancements, could they?

[Argus Collingwood]

---

Void Singer wrote:

---

Lydia Craig wrote:

[...] Two it would be nice if they would tell us what they intend to accomplish by this move, what we should expect while its it going on, and when servers are down for unusual times give us reports on their current condition.  Finally, it would be extermely good if LL provided a post event team to help with any problems that may arise during this process.[...]

---

they did tell you... you just didn't understand.... OS upgrades... which is just updating the machines that SL runs on, not SL itself (security patches, maybe some minor improvements on how they work), and what's going to happen (extra region starts, caused by them swapping out machines that those regions are running on).

since restarts are all that should be happening, regions shouldn't be down for excessive amounts of time. if they are, treat it the same way you treat any restart that doesn't come back up... report it to live chat, or via a ticket (on of the few ticket types that gets almost immediate response). and if something goes horribly wrong they'll post it in the gridstatus page as usual (not that I think that's effective or sufficient for a great many things, but it's what's likely to happen)

 

I do prefer the popup notices for restarts myself too though.

---

Server farm movement ya think? ;-) I prefer popups too.

[Void Singer]

nah, those could be pretty seamless, resulting in a single restart.

@general

I was NOT aware that the OS upgrades should affect the timewarp problem (but it was a hope in the back of my mind that the updates might improve communications between regions) but in general, OS upgrades are perfectly normal in the cycle of business... with single websites they're nearly unnoticeable, but with SL you're talking thousands of machines that have constant realtime streams.... you really cant swap those out without interruption. if they are non critical you can usually get away with waiting until the next machine restart and peg them in when the machine was going to be down anyways.

and when they do need to be done en masse, you can't just conjure up enough machines to do a straight 1:1 machine swap to make it a single mostly transparent restart with a network as large as SL.

[Chelsea Malibu]

Many of you may not recall when SL went offline every Tuesday for most of the morning to do updates.  I love that we don't have that now and as Philip said himself...Real Life doesnt shut down for updates.

Just think how boring SL would be without technical issues or the drama...lol.  It's just another sun filled day and a full moon night in SL.