Rolling Restart Tue/Wed/Thu June 24-26 to deploy 1.22.4
Tuesday, June 24th, 2008 at 11:46 AM by: Prospero LindenWe will be deploying Server Version 1.22.4 in a rolling restart this week, following our usual schedule:
- Tuesday evening : a “pilot roll” to ~200 regions
- Wednesday 05:00-09:00 am : a rolling restart to half of the hosts on the grid
- Thursday 05:00-09:00 am : a rolling restart of the rest of the grid
This server version includes some internal security and bug fixes, and should have no resident-facing differences from the previous version (1.22.3). Server version 1.22.4 is currently running on the “Second Life Beta Server” channel of the Preview Grid; for more information, see this status blog post from last week announcing that deploy to the Preview Grid.
As with all of our rolling restarts, each region should only be down for about 10 minutes. If your region stays down for more than 30 minutes, please contact support. All regions will receive a warning starting 5 minutes before they are restarted; there is no way for estate managers to delay the restart of a given region.
Update 2008-06-24 01:09pm: This is just a server update; no new viewer download will be required as a result of this rolling restart.
Update 2008-06-24 09:50pm: The rolling restart to about 300 regions is complete.
Update 2008-06-25 05:37am: The first-half rolling restart began at about 05:20.
Update 2008-06-25 06:23am: The first-half rolling restart is about 35% done.
Update 2008-06-25 07:17am: The first-half rolling restart is about 59% done.
Update 2008-06-25 08:26am: The first-half rolling restart is complete, but for a handful of down regions that we’re cleaning up right now.
Update 2008-06-26 05:15am: The second-half rolling restart is beginning now.
Update 2008-06-26 06:13am: The second-half rolling restart is 23% complete.
Update 2008-06-26 07:15am: The second-half rolling restart is 64% complete.
Update 2008-06-26 08:20am: The second-half rolling restart is done, except for a few reigons that are still down and that we’re working on right now.
June 24th, 2008 at 11:52 AM
Joy, I can already see all the stuck regions.
Rolling restarts tend to be nightmares.
June 24th, 2008 at 12:00 PM
Does this require I load a new viewer? Unlike some I enjoy every upgrade that gets done and look forwards to improvement,
June 24th, 2008 at 12:00 PM
I have 4 days i cant log
June 24th, 2008 at 12:01 PM
What does “resident-facing differences” mean? What bug fixes?
June 24th, 2008 at 12:10 PM
Oh rapture :/
June 24th, 2008 at 12:12 PM
Will this fix the bug where while in voice chat, everytime a bluenote comes across it cuts the volmue down to almost nothing. This happens not just to me but a lot of others.
June 24th, 2008 at 12:21 PM
More interested if this will fix…ha…fix avatar collision.
June 24th, 2008 at 12:25 PM
This is a server upgrade, it doesn’t affect what viewer u are using. No new viewer is needed. Resident-facing differences means that Lindens hope this won’t affect residents in any way. We shouldn’t, in theory, notice any difference in sl after the upgrade has been rolled out!!! However let’s wait and see!!
June 24th, 2008 at 12:27 PM
[...] blog.secondlife.com [...]
June 24th, 2008 at 12:51 PM
What are the bug fixes gona be? Or did I miss a post somewhere lol
June 24th, 2008 at 1:06 PM
Tristin@10: they’re not going to tell you what security fixes they’ve rolled out until after the rollout, lest people try and take advantage of the holes that they’re fixing before they’re fixed.
June 24th, 2008 at 1:09 PM
OMG!!! The sky is falling on our heads!!!
tbh the main problem will be the additional load for the asset servers during the whole restart procedure…
Let’s see if it can be handled this time…
June 24th, 2008 at 1:09 PM
SL is SH@te tonight rolling restart or not, but there again no real change, samr problems rejoined by all the old supposedly resolved ones too, Happy Freaking birthday SL
June 24th, 2008 at 1:10 PM
Any chance we’re getting SVC-2485 with this, Prospero?
June 24th, 2008 at 1:16 PM
So… it’s not really true that rolling restarts generate additional load on the asset servers. Sometimes rolling restarts have been rocky, although we’ve improved the process that at this point the vast majority of regions really are only down a few minutes. There was a bug in the server code a couple of revisions back that made the rolling restart to deploy 1.22.2 very unpleasant, but that was due to the bug that was being fixed. It had to do with a memory leak in the simulator, and had nothing to do with the asset servers. This rolling restart should go just as smoothly as last week’s.
Chief Blackhawk: What is a “bluenote”? In any event, this server deploy doesn’t change any of the Voice code; and, indeed, the low-level handling of Voice is done by servers maintained by Vivox. If there is a bug associated with Voice, please file a SVR issue using our issue tracker. As stated above, the changes to this server version are internal bug fixes, and nothing that changes the outward behavior noticably.
Jessicka: the avatar collision bugs are not fixed in this release, although we’re aware of them, and are internally testing code that fixes them. Those fixes should come out in an upcoming rolling restart.
June 24th, 2008 at 1:23 PM
Will LL sometime correct their IM´s to Email?… i know already 4 , they decline them as Spam!!!!
Return-Path:
X-Flags: 1001
Delivered-To: GMX delivery to anny_helsinki@gmx.de
Received: (qmail invoked by alias); 24 Jun 2008 20:18:34 -0000
Received: from nat.janus.lindenlab.com (EHLO web-ds12.agni.lindenlab.com) [64.154.223.153]
by mx0.gmx.net (mx060) with SMTP; 24 Jun 2008 22:18:34 +0200
Received: by web-ds12.agni.lindenlab.com (Postfix, from userid 33)
id 43585698444; Tue, 24 Jun 2008 13:18:32 -0700 (PDT)
To: “Anny Helsinki”
Subject: *** GMX Spamverdacht *** Second Life - the object ‘Eliz Watanabe Controller’ has sent you a message
From: “Eliz Watanabe Controller”
Message-Id:
Date: Tue, 24 Jun 2008 13:18:32 -0700 (PDT)
X-GMX-Antivirus: -1 (not scanned, may not use virus scanner)
X-GMX-Antispam: 3 (different HELO strings used by IP)
X-GMX-UID: rWBSfKs2TXs8dVcnwWU584ZKRzdyMsPS
CORRECT THAT, THAT IS TOLD U SINCE 2 WEEKS NOW!!!!
CANT BE THAT COMPLICATE!!!
TY
ANNY
June 24th, 2008 at 1:24 PM
To all those decrying the rolling restarts, would you prefer they went back to shutting down the grid for a few hours during a “low population” time of day like they did a year and a half to two years ago? The largest complaint of the time was from people outside of the US, that the down time happened during their only free times (after work/school) and lasted just long enough for bed time to roll around for them.
June 24th, 2008 at 1:27 PM
rETURN PATH IS Return-Path:
IS ONLY ONE OF THE WRONG SET SERVERS CORRECT THEM , NOT MY POSTS
June 24th, 2008 at 1:28 PM
OH,….. it is forbidden to sho up what massive failures you do.. okay i will shut up another week….
Return-Path:
June 24th, 2008 at 1:29 PM
Prospero, I think by bluenote they mean an Estate Message (ie, the blue notice appears in the upper right corner of the screen). I don’t use voice much but I did observe this myself several revs ago, and didn’t have a recreate-able info source to bugfile it myself. Haven’t tested it lately so I’m not a good piece of data.
June 24th, 2008 at 1:30 PM
yes feynt, was better to calculate for events. so 8 hours downtime on a monday will always be better than downtimes all the time in eventtimes and on weekends.
dear pretty sweet Lindens, please gibe us back 2 days of downtime a week we can plan with and not ur coffee-times as downtimes when u play pool billard on screen or delete Return-Path:
June 24th, 2008 at 1:33 PM
btw, Feynt, minoryties shouldnt be able to tell the mayorities what to do, its called Democracy, only if you never learned that….
LL said more than 60% are outside the usa… questions?
June 24th, 2008 at 1:46 PM
If you own land and restrict access, why?
For private personal use
For confidential business use
I don’t own land
Other
will people without info of CC or paybal be banned after answering this ?
hell, Katt, what is that complicate to ask questions why people DONT PAY LL 10 BUCKS A MONTH FOR THE INFORMATIONS YOU WOULd LIKE TO HAVE?
Here is the answer, they dont trust LL!!
the Age-thing is the right thing to believe you would sell all our data.
btw it still didnt work for countries that secures their residents from data-robbers.
and now i try to go online
June 24th, 2008 at 1:54 PM
I would much prefer SL be down for a day or two, as they massively take on all bugs, confirm that they are fixed through several tests, release a new viewer the provides most of the significant bugs fixed, no “known bugs”, and go from there, versus this system of try and fail, and fail again, and fail again, and fail again, and fail again.
June 24th, 2008 at 2:09 PM
Yes I would rather it be down for 5 hours . This way I don’t have to pay someone to stand at my region all day waiting for a restart that MIGHT happen today, tomorrow. or the next day …, Not knowing when your region is going to be restarted and not knowing if it is going to recover and come back up is costly .
Everything is so vague when it comes to restarts. and you only know 5 minutes ahead of time
Going to the preview grid when SL was down was really a lot of fun and everyone was able to relax and take a break for a bit.. Ahhhh the good old days!
June 24th, 2008 at 2:19 PM
a sort of related question..i heard it was possibly to build above 756m but i’m running 1.19.1 (4) and obviously cant, when does the new viewer come out thats passed its beta testing phase?
June 24th, 2008 at 2:32 PM
Does this include a fix for http://jira.secondlife.com/browse/SVC-2511
PLEASE say it does !
June 24th, 2008 at 2:36 PM
ok i just saw above and saw that they are not fixed
/me drops head and sighs………….
June 24th, 2008 at 2:53 PM
i would just like LL to fix or stop inducing the resetting of show in search/for sale prims to not for sale, not show in search, and especially stop changing things to free to copy. This type of data manipulation needs to be logged, analyzed, and dealt with. If it is a person without skills in the database then change the password. If it is a defect then fix it and keep it fixed. Does LL even have a clue as to what it means when they roll this defect back out? It means merchants have to go check every single prim all day long. May as well close SL if they can’t fix and keep things like this fixed.
June 24th, 2008 at 2:59 PM
>>Does this include a fix for
>>http://jira.secondlife.com/browse/SVC-2511
hunh?
JIRA, much like the Wiki, is unreachable for two months. Day or Night. Doesn’t matter. Browser just hangs forever on “waiting for data from jira.secondlife.com”
I’d have liked to report a couple of reproducible bugs from the RC but with JIRA unreachable for mere mortals….much like the other ongoing problems these too are apparently going unnoticed.
RH
June 24th, 2008 at 3:14 PM
OMG! Everyone else is in a panic, so I better panic too!
Help!
We’ll all be murdered in our beds!
‘Jessicka Graves Says:
June 24th, 2008 at 1:54 PM
I would much prefer SL be down for a day or two’
What? Ten minutes per region is ok with me… A day or two? That would mean having to venture out into the real world again… ughh!
June 24th, 2008 at 3:29 PM
Are issues with buoyancy going to be fixed soon ?!?
(SVC-1792 : llSetBuoyancy bugged in Havok4 - says a fix is pending but it’s been like that for months!)
(SVC-2013 : Hovering avatars sink Slowly since Havok 4 Rollout - only partially fixed in 1.22.2; you still sink if you turn!)
June 24th, 2008 at 3:58 PM
I think what is trying to be said is that limiting restarts to specific days between specific hours is much easier to plan for then having them happen at any time. I, as a business owner, may plan an event months in advance. Spend thousands of $L for staff, entertainment, advertisements, and accessories, just to have LL decide a week before that their going to restart my region 15 minutes into the start of the event, with the possibility, as has been demonstrated in the past, that the region may not recover for hours. Not to mention what ever new bugs that has been introduced by the update itself. Many large IT based corporations have what is called maintenance windows. These types of restarts are restricted to happen only in those windows, and require user approval to happen outside them. Even in the case of an emergency, advanced notification is required, as well as top management approval. When these emergencies happen then extensive root cause analysis is required followed by a comprehensive plan to prevent it from happening again. This is not a new concept but IT 101.
June 24th, 2008 at 4:03 PM
so that explains my sudden lag issues I have been having today. Well a little lag while the resart rolls through in exchange for fixing the problems is a good trade off.
@Anny Helsinki Why should anyone even bother with becoming age verified? Who can name one single place with adult content that is flagged? Like before any one old enough to turn on the computer can access these places. i would not even waste your time becoming verified, LL doesn’t enforce it any how.
June 24th, 2008 at 4:17 PM
My guess is that the rolling restarts seem to affect nearby regions as they try to re-establish communications and bandwidty usage on the network goes up. But I would much rather see a little lag vs the older method of just shutting the whole place down until the deploy was done.
Being one to often take LL to task, I find it ironic that I feel the need to stand up for the process now, but folks, you gripe about issues, you gripe about the grid being down, and then when LL roles out fixes without taking the grid down, that gets griped about too.
Yes issues are disruptive, yes there definately could be much better communications, yes rolling restarts can be a pain. But if we want things fixed, we want a more stable grid, we have to give someplace.
Havoc4 has cost me a lot of time redoing things that no longer work, or to remove hacks that are no longer needed, but I happly trade off that time for a much more stable world.
Good work on fixing the holes Lindens, and on rolling restarts vs a closed grid. Cherry pie for all…
This time.
DRD
June 24th, 2008 at 4:27 PM
@35 uh-oh, I agree on the age thing, if anyone really wants adult material, can just google it, and turn off the filter.
@34 RichD, I applaud anyone who speaks common sense like that, I don’t know how many times I’ve been in the middle of something important and gotten taken out by a rolling restart.
I finally gave up on running a business in SL because of the glitches, and this sort of no actual warning about restarts. Just telling us that its going to happen on Tuesday or Wednesday isn’t good enough. We the business owners don’t feel LL provides support on this.
Between the glitches, the lag, and never knowing what LL is going to do next, I just said f*** it and gave up. I don’t have the money to fight all that and wait for it to be fixed.
My best wishes to those who are willing to try and hang in there, but I’m not wasting another dime on trying to run a business in LL.
June 24th, 2008 at 4:49 PM
I don’t even own land, but I agree: more advanced notice, say 24 hours, of a restart would help out landowners enormously. That way they can make whatever arrangements they need to in order to have someone present to monitor the situation, postpone/cancel events, or anything else they need to do.
June 24th, 2008 at 4:50 PM
midnyte@26: you have to use the 1.20 release candidate to build over 768 meters. I have asked a couple of lindens why they couldn’t release a new 1.19 version that has the 768 meter limits removed, but they don’t consider that a feature of Havok 4, so it doesn’t need an update to the viewer.
I believe some of the alternate (open source) viewers support it as well.
June 24th, 2008 at 4:54 PM
I never paid attention but do openspace sims return to the exact same host after a rolling restart? I guess I’ll find out in the next couple of days. Would be nice to stay where I’m currently at.
June 24th, 2008 at 5:04 PM
@Anny, you might want to complain to your ISP email service first.
“X-GMX-Antispam: 3 (different HELO strings used by IP)”
Tell them using HELO to evaluate email and rate spam is not recommended. See the RFC’s. It is not a requirement that the HELO matches the IP domain.
Return path is not required, and from what I can tell, SL’s mailer is not putting any return path in the header, that is your ISP screwing up .
June 24th, 2008 at 5:09 PM
I wish you luck with this update and hope it goes smoothly. Any time frame when Mono will be deployed to the main grid?. Been watching people working on it in the preview grid and wow can`t wait.
Good luck!
June 24th, 2008 at 6:02 PM
Nothing will change I suppose that means the 10 to 12 second delay when opening a second inventory window and the 5 second delay when opening a script. It seems to me that this has developed since you have been playing with the latest versions. Inventory windows used to pop immediately as did the script window. In another vein are you people ever going to fix sculpted prims so that one does not have to right click them to make them resolve. And yes it happens on any size map from 64 x 64 to 256 x 256 also there are some that never make it to the cache. Again you have so many problems I am surprised anything ever works in this game. The 5 year (9) birthday and it still functions like a beta test.
June 24th, 2008 at 6:13 PM
#43 coventina dalgleish
you might want to update your graphics driver. I know with Nvidia the last update was a flop and they released a new one the 23 of this month .. I used to have the delays as well and it was all because of my video driver.. Since updating to the newest version All my interface windows Open instantly..
Something for you to look into, of course I do not know your configuration but can sure share with you what fixed my problem with interface window delays..
June 24th, 2008 at 7:03 PM
**I don’t even own land, but I agree: more advanced notice, say 24 hours, of a restart would help out landowners enormously.**
Please take a look at our Status Blog. We try to keep the tentative and firm dates of rolling restarts listed in the Google calendar which is excerpted there. Tomorrow and Thursday’s restarts have been listed since at least yesterday, and I believe since last week.
Please note also that the rolling restart windows Wednesday and Thursday mornings are five-hour windows. Rather than tell you a five hour period when the entire grid is down, we are giving you the five hour periods sometime during which each region will get restarted.
Re: do regions return to the same host: normally, they do not necessarily. Each time a region crashes or restarts, whichever host is available will run that region. However, during a rolling restart, we lock regions to hosts, so that after the planned restart they will go back on to the same host.
Re: Mono, current plans have it in likely to be part of 1.22.5. We will most likely roll out 1.22.4 in two weeks– that’s to clean up everything else that’s being done before we merge in Mono. Then Mono would come out two weeks after that. This might slip if there are any critical security bugs that require us to do a second rollout after 1.22.4. You will also require an updated client to support compiling scripts for Mono; it is possible that an updated RC with that capability won’t come out until the week after the Mono server hits… or, it could come before. I’m not sure at the moment.
June 24th, 2008 at 7:03 PM
CORRECTION : current plans have **1.23** in two weeks, and Mono as part of **1.24** two weeks after that.
June 24th, 2008 at 7:26 PM
Can we get the % done again? and has todays ~200 started yet?
June 24th, 2008 at 7:34 PM
Thanks Prospero for the replies.
June 24th, 2008 at 7:36 PM
Keep up the great work Prospero! Rolling Restarts are much better than it was before when the entire Grid was down on those dreaded “Update Wednesdays.”
Anxiously awaits Mono!
June 24th, 2008 at 8:16 PM
Anny:
darf ich vorschlagen, dass Du Dich mit GMX in Verbindung setzt - Linden Labs hat mit der Tatsache, dass deren Emails als Spam unterschlagen werden, nichts zu tun. GMX ist aeusserst aggressiv, was Spam angeht - ich weiss das aus eigener Erfahrung, da meine Eltern in Deutschland GMX benutzen und ich vor ein oder zwei Jahren nicht in der Lage war, ihnen Email (von meinem Roadrunner-Account in den USA) zu senden… und das lag daran, dass GMX meine Adresse als Spam-Relay auf die schwarze Liste gesetzt hatte. o_O
To those who don’t speak German: don’t worry, you didn’t miss anything. *grin*
June 24th, 2008 at 8:28 PM
So does this sort of stuff happen all the time?
Just joined but this thing keeps telling me invalid name/password and nobody takes any notice of my queries. I use e-mail OK and buy stuff in various places OK but this thing here simply does not work. Why should this be so? How can you all use a thing that seems to carry permanently a fault condition?
You all seem to think that it is worth while persisting in trying to make it work. Has any one had success with it yet?
Please do not refer me to the support page.
June 24th, 2008 at 8:29 PM
Will this fix the constant crashes I seem to be experiencing lately? I’ll give you an example of today’s performance. I first logged on - lasted just over an hour before crashing. After that? 2 minutes, 2 minutes, five minutes, fourteen minutes, four minutes. And you call that a stable grid? And to the first person who says it’s my computer - go jump. I’m running -admittedly with a slightly older video card but am combining that with 2 x 2.8 ghz intel processors - more than twice what is required to run sl. I also have my Second Life settings on absolute minimum to try and get some form of stability. This results in abhorent rez time, but the stability is still shocking. Why don’t Linden labs get with the real world and use reliable asset servers (like all the other metaverse companies do) instead of supporting some backyard fix-up job? And why - when everyone who works at Linden Labs is suppose to be oh-so highly skilled and intelligent - can they not have this thing stable after five years? I’ll tell you why. They are more interested in getting the money out of people than providing a stable gaming platform. They cater for those few of us who can afford to update every time Linden Labs do a new Regular Client and who insist on brown-nosing. In australia we call that a “liberal attitude” - give to the rich and screw everyone else over. Linden Labs do not care about anyone but themselves - so long as they get money and there are a minority of people who will kiss their A**.
How you going to respond to that Prospero?
June 24th, 2008 at 8:35 PM
By the way, happy fifth birthday, Second Life. If this were a real child, Linden labs would have been reported to welfare years ago for the maltreatment they have inflicted on second life.
June 24th, 2008 at 8:50 PM
I’m very excited about Mono, hope everything works out well. If scripts will really load at least one times faster, oh boy, that will be a major step forward.
Still wrestling with havok4 though. Swimmers and hugs, will there be a fix?
June 24th, 2008 at 8:58 PM
do you always use the same regions to do the first roll out?
June 24th, 2008 at 8:58 PM
/me facepaw dee, I spit coke out my nose there lol
June 24th, 2008 at 8:59 PM
55@christos: I beleive that it changes periodically. I dont remember exactly but I do know it does not roll from north to south anymore. (miss thost too I do)
June 24th, 2008 at 9:02 PM
SLS 1.22.4 rolling restart starting tonight, unspecified fixes…
There are no listed changes for this rollout, which experience tells us mean that all the changes are security/exploit fixes. Some users are a mite testy with the lack of advance warning. Prospero Linden says that advance notice of the rolling restart …
June 24th, 2008 at 9:09 PM
@44 Beerandpizza
Thanks for the tip - Got the latest and it definately does help with texture and sculpty loads times.
@52 Dee Tunwarm
Just jumped, now your turn - Your connection or machine are most likely at fault. Video it the biggie here. I have SL running on an old Dell P3 with a NVidia 6600 AGI and it runs just fine. I have not had any real stability probs in some time now. Memory, more Memory and the best Vid you can afford, thats the key.
I have been in here two years now, and I can guarentee that the majority of lag and restarts are local problems, not SL problems. Yes SL has issues, but if you are the only one around crashing, or the only one suffering lag, guess what… its probably not SL.
DRD
June 24th, 2008 at 9:11 PM
BTW, that P3 is not my main machine, I run twin 7600’s with SLI, 3 gigs, AMD Dual at 3900. The only issues I have are with my dumb cable company.
DRD
June 24th, 2008 at 9:26 PM
Well said, DRD! Hi-five to ya!
June 24th, 2008 at 10:16 PM
Prospero -
A wiki page with “a look at the next month” even if the dates were ‘in the next two weeks’ would be helpful to see whats coming down the pipe. Any chance such page could be created/updated by someone? Or even a notification in SLDev mail?
June 24th, 2008 at 10:39 PM
G’day Prospero, ( like the rocket name)
This may be the wrong spot but did something happen with windlight? I dont get the oooo ahhh views like I used to do. I haven’t made any changes to my graphix settings , in fact there maxed out. But SL is still looking a touch drab.
Perhaps I missed a anouncement .. Oh and Im using the latest vanillia viewer (from SL).
June 24th, 2008 at 10:39 PM
I noticed when I lock up, the problem is bandwidth. (Ctrl-Shift-1) It goes down to 0 and I can’t do anything until the bandwidth pops back up.
I am using DSL. I wish there was a way to reset the internet connection without quitting and starting back up.
June 24th, 2008 at 10:41 PM
@Prospero
You’re updating the live grid with Mono in a few weeks? Will you be releasing more information about how the transfer will take place closer to the date?
Just wondering if it will be ‘dual’ like the Beta grid is now or if it will replace the existing engine ‘transparently’
June 24th, 2008 at 11:50 PM
Really, what happened to the calendar of support events? I can’t find it! It used to be that I could look up for the next scheduled maintenance on that calendar and make the plans accordingly. But now, where is the link? I don’t see it on the main web site anymore.
June 25th, 2008 at 12:04 AM
Mono will not replace the old LSL2 scripting engine. It will run alongside and the user will have the option to compile their scripts using Mono or LSL2. They said they might disable compiling for LSL2 in the future, but for now they will be seperate engines.
June 25th, 2008 at 12:16 AM
i have to agree with the girl complaining about the gmx issue.
the past week or two all incoming mail is being interpreted as spam due to changes ll implemented in the mail delivery. adding the ll domains to the whitelist has no effect, as the spam trigger appears to be:
X-GMX-Antispam: 3 (different HELO strings used by IP)
for once, just once, please do something that makes me happy and fix this!
June 25th, 2008 at 12:35 AM
Okay, I’ve been frustrated with SL downtime in the past, but now people are overreacting…
There are enough ways to be informed about scheduled SL downtimes, besides it’s only 10 minutes per region.
Is it really that hard to check status.secondlifegrid.net now and then? Heck, they even have a Google Calendar which lists the upcoming interventions. It’s iCal compatible, so it should integrate with most popular calendar tools too. I’ve added it to my personal Google Calendar, and I was aware of the restart for a few days now …
Not good enough for you? Okay, use the RSS feed then. Firefox has an integrated RSS newsreader (I think IE7 has one too).
June 25th, 2008 at 12:51 AM
Ha ha @ DR Dahlgren…I have the fastest MB intel makes, the fastest chips, and running 2gigs of DDR3 memory, which btw is the most Vista can really use unless you run the 64 bit version, and I have the 8800 GT Nvidea video card maxed with memory. I get the same lag everyone else gets, and yes, I have a cable connection with avg 800 mbs speed. I have friends running monitors on SL memory usage, and it is unbelievable how it can go from around 600K to over 1.5 gigs very quickly when the memory leak starts….something still wrong there, though it supposedly was fixed. Of course I stay online more than most maybe, 12 hours a day and I crash a few times from the memory thing, and hit periods of “freezeframe” that last quite a while. Even on private estates with only 10 avatars its bad. heck I can be the only one on my island and have lag. so apparently being on your own server doesn’t prevent SL-wide lag from affecting you.
But I’m addicted like everyone else, so I just hope they get it this time. Ive only been here just over a year, but I remember it being better with my lowly laptop when I started as far as lag. In fact I never even heard of it until after the great crash of May 07 and it has not been the same since. There are several viewers 3rd party that only use around 500K instead of the 900 the LL one needs to run on…and they work much faster..in fact a lot of the lag is caused by your system being short on memory! RELOG RELOG RELOG.
June 25th, 2008 at 1:27 AM
Dear Ron, it isnt me or GMX that is doing a failure.
LL need to speak to them and change their “helo” on some servers.
only from SOME messages i got it as Spam, most other works fine, because they send their IM´s from another server.
and please, dear Ron, look at post #68 too, its not only me :))
ty for reading me
nice greetings from
Anny
ps: most what is disturbing me, is i receice all spam now, because i need to open the spam-filter to receive all messages from land and friends in SL.
June 25th, 2008 at 1:52 AM
*yay Lindens
June 25th, 2008 at 2:37 AM
[...] Source: blog.secondlife.com [...]
June 25th, 2008 at 3:03 AM
@ 70 Nikolai Kidd
No argument on the mem leak issue. If you go into an area like a mall with a ton of single item vendors and all those pictures, you can kiss your mem goodbye. But that is still a client side issue, not a server or stability problem, which is what was being discussed.
I would seriously watch the sim stats when you see lag and are the only one around. If they look good, it almost has to be a connection issue.
As to lag, I can not remember a time with out it, but def not the problem I remember it being, back when 15K ppl in the world brought the whole place to its knees.
Yes, SL still has problems. I doubt that it is realistic to imagine there will ever be a time when it doesn’t. Why, because it will never be finished, they will never quit trying to add to it, or improve it, or change it.
And for that, I am grateful, especially when we can get them to listen. Thanks for the skins on the RC LL.
DRD
June 25th, 2008 at 3:32 AM
Prospero with all due respect those status reports are usless to land owners if they still have to wait for 3 days and guess when their region my be restarted..
Most ISP providers can tell you What time and date your sever will be down . We all have to sit around and guess and waste time and money paying people to hang at their sim to make sure it comes back Up. Otherwise it will stay down and never be taken care of..
Some It is extremely annoying and costly . And when someone forgets to update the status of Restarts we end up having our help hang around longer than needed. You all know which clusters you plan on doing and some decent information on which cluster we all are on would help and would let us know what time and date .. Instead of a 3 day guessing game
June 25th, 2008 at 3:40 AM
@75, have you even botherd to read the comments hear, ill vouch for this i know for a fact that the rolling restarts were on the status page 7 days ago. its not hard to bookmark a site so stop complaning that you got no warning when the only reason you dident is because u never checked
June 25th, 2008 at 3:49 AM
Yea i read it and still 3 day guessing game .. will it be restarted today? or tomorrow? we dont know and have to twiddle out thumbs waiting instead of having denet info when it will take place. I Am fully aware of the warnings but get real 3 days to guess about it is rediculous That still does not tell you which day your region will be restarted . Only a 3 day guessing game . do you know exactly which day your region will be down ? No you dont and that is the point i am trying to make..
Yes by reading it still doesnt tell me if today is the day for my region . Whoops could be today or could be tomorrow.. yes guess which day it is???
June 25th, 2008 at 3:58 AM
Please don’t use blog comments to ask help about unrelated things - This is not a support forum.
You may usually find better help in the webforum of your choice.
June 25th, 2008 at 4:16 AM
Dilbert,
Don’t worry about it. Afterall, LindenLabs are “pioneers” and they are doing “revolutionary stuff”, so you should accept those issues…
That are your own words from a previous blogpost, quote :
–
There are many bugs and issues they have no clue how to resolve. I guess that all comes with the one and only pioneers who are paving the way right?? When the world began Way way back before 5,000 B.C. there were many bugs and issues the humans had to learn and endure. Like mass diseases and deaths and no clue how to stop it.. But as time went on humans worked hard and pressed on and now today people live longer and have better lives than way back when .. It is all a part of evolution . As this world grows the same style the pioneers will pave the way and things will be resolved.. LL has set the stage and are the learners in this as well. you can not expect all of it to work perfectly but you can strive and work toward making it better.. If everyone in the world sat on their buttas complaining al the time and not working t=o find a solution to problems . then the world would not be as it is today and you would not be here to complain .. It is evolution
—
Seriously, make up your mind and stop contradicting yourself …
Sartek.
June 25th, 2008 at 4:29 AM
Lol yea Im sure you guys love the flaming . My point a while back was not really about complaints but the flaming Of Linden labs. I raised a valid point and Have fun Posting against others posts .. Even tho what i have said in the past you do not understand my point on the flaming of LL and complaints that are valid are good Because they value free expession and Yes many bugs are out there and have to deal with it daily. And have realized after being on SL for 2 years thats how it will be today tomorrow and the next day for bugs I merely raised a valid point and not posted against you or anyone else. Flaming LL is not my game . I only ask for some decent info Thats all . Have a good one Sartek . Maybe you could post about other things instead of trying to flame users and LL. I have made up my mind to deal with it and have been by continuing to use SL .. And thats what i have done . made up my mind weather or not i can take it.. I respect everyones post but when the flaming gets out of controll Yea it looks sickening in here > My post was valid and productive . Needless to say about yours lol
Have a good one Sartek .
June 25th, 2008 at 4:31 AM
No new viewer? that’s a shame :-/
June 25th, 2008 at 5:04 AM
Is a server upgrade necessary every week?
Why not monthly and then take the grid down on a fixed moment,
so we know when and why. Let’s say 1st Tuesday of the each month.
June 25th, 2008 at 5:32 AM
Dilbert,
It was not intented as a flame. I wast just surprised that you make a big deal about this 10-minute downtime of a region which was communicated well in advance. When we complain about a not-so-well announced downtime that took a lot longer, you accused us of flaming LindenLabs etc…
So yes, I think my post was justified.
June 25th, 2008 at 5:36 AM
@82 Linda Brynner
If there are security problems, waiting on average 2 weeks to fix it is not a good way to handle it. In this case it *is* a security fix and I really want such things to be fixed as soon as possible.
When it comes to adding new features and/or functionality, it’s always a compromise between the plans made for the future and the experience of the users. With an aggressive deliveryplan you’ll get updates with short delays, once per week for example, and that seems to be LLs plan at the moment.
Keeping it to once per month, unless it’s a security fix, might be a good thing when the product is stable but that’s not the case at the moment.
June 25th, 2008 at 5:39 AM
I don’t really see the problem with rolling restarts at all. It’s true, you don’t reall yknow when your region will be restarted, but it’s a very short downtime anyway. I rather have this 10 minute downtime, during which I can use most of the remaining grid normally, than have the full Wednesday grid outage back we had a long time ago for each update.
Having an exact time available, as to when your region will be restarted would improve this process of course, but I really don’t see it as a big deal, even less worth to make such a fuss abouot it.
June 25th, 2008 at 6:14 AM
@69 Sartek: boy, you speak from my heart. I’m 100% totally with you. Hear, hear!
@71 Anny: point taken, I guess your issue is different than the one I had with GMX. Hope it gets resolved, though (for the other person as well), so good luck to you
@86 Zi: again, I agree completely with you - the way rolling restarts work now is better than what we used to have. Surely anyone is able to teleport with a 5-minute warning?
June 25th, 2008 at 6:35 AM
LInda Brynner and aSwede — aSwede mostly answers your questions, and he is right about there being a tension between wanting not to disturb the grid and wanting to get upgrades out there.
One point, however: currently, we *don’t* upgrade the server every week. We didn’t last week. We won’t next week *unless* there are serious security or other issues that necessitate a rolling restart. We did have several weeks in a row where we did rolling restarts because we had “ZOMG gotta get this out or we’re gonna die” patches that kept cascading on top of each other. When those things are around, we really need to do a server upgrade.
As I said, the default plan is that in two weeks we’ll do a larger update (1.23) that will include a number of fixes and changes to get ready for Mono, and then two weeks after that we’ll do a large upgrade (1.24) that will include the introduction of Mono. Our current process has any new major server upgrade baking on the Preview Grid for at least several days, ideally a week, before we do the rolling restart to the main grid. We *don’t* always do that for the “ZOMG right now” patches, because those by their very nature are things that *need* to get out.
The danger of waiting longer between subsequent upgrades is that there are more changes in a given upgrade, and consequently a greater chance that an undiscovered bug will introduce instability. Yes, we do QA, and yes, we test things on the Preview Grid, but the fact is that the scale of the main Second Life environment is well beyond the scale of any testing environment available. This means that there are two things that we might not catch in testing; first, scaling problems that only show up with the full load of the main grid is running the new server, and second, very rare edge cases, or specific functionality that only shows up when a *lot* of people are doing the diverse things that people do in Second Life. As such, it becomes advisable to try to limit the number of changes you make at once, so that you can have a better handle on them. Thus, most often we do server updates every couple of weeks instead of doing a huge server update every month or two.
June 25th, 2008 at 7:13 AM
Been waiting aaaaaaaaages NOW !
for a SL mapview update !
cryophile sim has not had a map texture update for at least 3 months !
suckage !
like most crp here !
June 25th, 2008 at 7:25 AM
Dear Prospero,
Everyone who has a business or is planning an event wants to know when *their* sim will be restarted, either to make sure it comes back up, or to adjust the event time to avoid an interruption. So it would be massively helpful if you tell us what method you use to sequence the restarts.
Ideally that would be something simple that we can check ourselves, like all odd numbered servers in server number order today, and even numbers tomorrow. For those who don’t know, the server number can be found in world in the Help/About second life item, where it says “you are at…..sim2345.agni.lindenlab.com”. 2345 would be the server number.
If the restart sequence is something opaque to us users but easy for you to do (such as by rack and card in the server room), consider that you are making life easier for the one or two people running the restart, and harder for the hundreds or thousands of people affected by business and event interruptions.
Even just normal players in a normal situation would like to know *where* to go when the 5 minute warning popup appears.
June 25th, 2008 at 7:31 AM
The only Sartek
Dilbert’s post was not even close to to a problem with 10 minute down time. If you would carefully read his post he clearly Asks for better communication on which day out of the 3 his region will be re-started… It is amazing how people misread things. I am all for letting us now the exact date . And not a 3 day maybe guessing game.. maybe Tuesday. Wednesday or Thursday?
He didn’t say 1 word about being down 10 minutes .I am sure he would be happy that it was only 10 minutes and not 3 hours like it has been in the past..
Kudos Dil you rock lol
June 25th, 2008 at 7:52 AM
Grip, Grip and more Gripping. All those who just like to complain. Please go back to RL and say bye bye to SL. If all you want to do is b*t*h then leave! It’s very simple. If you don’t like your job, find another. If you don’t like your dentist, well find another. If you don’t like SL then, well try and find another vritual world that is like SL.
Good luck!
I have problems like everyone else in SL. But the joy I get out of spending time there outweighs the idea I have to complain. I just file a support ticket and go on with my SL the best I can.
If it’s no fun.. Then leave.. make sure to shut the SL door when you leave and don’t slam it behind you.. there are folks sleeping
June 25th, 2008 at 7:53 AM
As for exactly when *your* sim is going to be restarted: that’s a challenge. Today I’m doing the odd-numbered hosts, tomorrow the even-numbered hosts. However, if I had told you that on Monday, it might not have been useful. Every time a region restarts, it comes up on a new host. As such, you might check Monday, see you’re on an odd numbered host, and plan for the restart Wednesday… only to find out Thursday that your region had in the mean time crashed or otherwise been restarted, and was now on an even numbered host.
Even today, knowing you’re on an odd numbered host, I can’t tell you exactly when the restart will hit. Yeah, they go in order, but things go slower or faster based on how loaded each machine is when it tries to shut down, whether or not I speed up or slow down the rolling restart based on the database load and other system issues, etc. It’s not done like clockwork and lockstep.
Why not release as much information as we have? The problem is that it could easily be construed as a false promise, as in the example above, and then people would get angry that we went back on the promise. Really using the information requires a fair amount of understanding, and the information without the understanding is *worse* than not having the information. What’s more, it could lead people to try to game things, e.g. by calling support and demanding that their region be moved to an even numbered host (not something that support can easily just do).
We are discussing and working on ways to give estate owners some more control over 4-hour windows in which their region will or will not be restarted during a rolling restart, but at the moment we don’t have that capability.
June 25th, 2008 at 8:03 AM
Thank you Prospero for explaining this..
I was under the Understanding that you locked regions to hosts during restarts But i guess that is only The current ones being restarted.. Even knowing that odd and evens are re-started on a given day Would be a big help also . And knowing the fact that if our region was restarted or crashed within that 3 days then we would be looking at a guessing game. Region Crashes are becoming a rarity . But thats much more info than we started with thats for sure.. Glad to see you all are discussing and working on ways to give estate owners some more control over 4-hour windows in which their region will or will not be restarted during a rolling restart
Any help in that matter would be appreciated by many I am sure
June 25th, 2008 at 8:40 AM
This SL thing might be nice and all, but I can’t even register at the moment…
“Your Second Life account could not be created at this time. Please try again in a few minutes.”
June 25th, 2008 at 9:15 AM
Would any of the timeless complainers please let do LL ops team do their job quietly ?
— thx — bored with all those 5 years old children minds, you better eat porridge and ask mom for a candy sweet you guys instead of whining endlessly. Enough !
And out of topic happy SL5B
to ALL LL employees, with a SPECIAL THINKING for ops and hot-line support that are most of the time agressed on that blog without any reasonable reason.
June 25th, 2008 at 9:20 AM
Sitting on an openspace sim watching the time dilation hover at 0.2 for extended periods of time. Assuming it is due to the other openspace sims on the same host going thru restart one at a time. Not a pretty sight!
June 25th, 2008 at 9:26 AM
Prospero, if you haven’t seen it yet/heard about it, there’s talk in the forums about scripts needing to be reset after this upgrade. Dunno how real it is, though.
June 25th, 2008 at 9:39 AM
@97 — (quote) “Sitting on an openspace sim watching the time dilation hover at 0.2 for extended periods of time” — Now detourning your wording, i do not wanna hurt but that makes me smile:
“Sitting on a cornflake
Waiting for the van to come
Corporation teashirt, stupid bloody tuesday
man you been a naughty boy
you let your face grow long
I am the eggman, they are the eggmen
I am the walrus
Goo goo goo joob”
http://www.youtube.com/watch?v=cqOKvonLrH8
Listen and take it easy, things will come back to normal on you sim — smile
June 25th, 2008 at 9:40 AM
@97 I take that back… time dilation is now stuck at 0.2 continuously. Time to submit a ticket.
June 25th, 2008 at 9:48 AM
@51 Slate
Why wouldn’t you go to the support website? You have a new account and can’t log in, right? They can FIX IT, unlike so many technical issues that SL seems to have. No one here in the blog is going to be able to help. Just call support and get it taken care of.
June 25th, 2008 at 9:54 AM
@100 — (quote) “Time to submit a ticket”
Yes please, but i suggest this one :
http://www.youtube.com/watch?v=etjpcF2X_mY
NOW LET THE SUPPORT AND