Rolling Restart to deploy 1.24 the week of 2008/08/18

I just wanted to get post #150
But I'm having fun in the sims that I visit that are mono enable, so things are looking pretty good.

<3

ty

My sim is upgraded to 1.24 but I'm experiencing the WORST lag at the moment. Cleared cache several times, restarted sim, checked my internet connection. All looks fine. And... I've found out that I need to reset scripts manually in order to get them to work. I am a furniture builder, so these are simple sit scripts amongst others. Is this Mono or just the worst lag I've seen in the last months?

If they're just poseballs and all have identical scripts then do the following.

1) Open the script up in the editor, tick 'mono' and save it
2) Copy that script into your inventory
3) Remove the script from all your other poseballs
4) Copy the script from your inventory into all your poseballs.

This will drastically reduce memory usage on the sim as the script is only loaded in once.

(You may want to wait until the rollout is complete if you're selling them! )


Desperately waiting for my friend's sim to be upgraded. I wrote her a game, but the script is dependent on mono due to memory usage and the need for faster calculations. I could split it up but then it'd be slower and use more link messages, which is bad for the sim, right?

Yeah, same thing here.
My land is on a sim running 1.24 (gotlib) and its unusable.

I would like to try and find out weather this following jira was intended or weather its bugged, http://jira.secondlife.com/browse/SVC-2927
us estate managers cant work like it D: please verifie!

I was on several 1.24.3 regions the last couple of hours and i had no lag. It just ran as smooth as always.

Thanks Lum. I understand that this option is only available in the RC, I'll try that. BTW, uploading animations isn't working either.

Ok, I have just discovered that my sim is now running the 1.24 server. It wasn't earlier today, at a time thatw as after the completion of today's deployment, so I was surprised. Pleasantly.

I am logged-in with the beta-grid client, and I have successfully compiled a few minor scripts to Mono. Here is my question: I have a number of objects with multiple scripts. For example, one has 27 prims, with maybe 1 script in most of the prims, and 3-4 scripts in the root prim. Is there any way to recompile all the scripts in the single object all at once? The wiki only describes compiling the scripts in multiple OBJECTS at once, and how they must all be selected at once to compile.

It should work the same, select the one object and recompile scripts in selection. If some of the scripts are identical, you might want to do it a little differently. Some of the memory savings come from copying around an already-compiled Mono script instead of compiling individually, so a little bit of juggling might be worthwhile.

Hrm.... I seem to be having a strange issue. Its not Mono-code related, but rather Mono-server related.

Essentially, scripts with XML-RPC functions seem to be failing under server version 1.24.3.95195. The same failure occurs regardless of whether these scripts are compiled as LSL2 or Mono.

The exact same set of scripts, compiled under LSL2 but running in an adjoining sim running version 1.23.4.93100 work perfectly fine.

Are there any known issues with XML-RPC in the new Server code? I seem to be having trouble opening a remote data channel, similar to SVC-2321/DEV-16211.

I'm still trying to debug on the RPC Server side, but something seems to be wrong, considering I'm getting completely different XML-RPC results on two adjoining sims running different versions.

Periapse/Babbage, if you'd like the scripts, I'm happy to drop them on you. Its pretty easy to repro the difference between regions.

Supposedly the 1.21 client wasn't supposed to be released until after the 1.24 server is grid wide.

My sim is 1.23 and I'm seeing the 1.21 client availible for download. What gives? This actually has me a little miffed, because I have a project going on my sim where we've been having to sit on our hands waiting for mono to come out. Now I have people coming in with 1.21 client and saying "Mono's out now, does that mean you have that new update ready?". To which I check the server version and it says 1.23 and end up having to explain to three dozen people a day why our scripts haven't been updated for mono yet.

The entire reasoning for releasing client 1.21 only after server 1.24 is grid wide is to save scripters and LL from the headache of having to answer questions as to why scripts work in some sims but not others. Now I'm feeling like I've been slapped in the face because I'm having to answer an entirely different question because 1.21 was released earlier than advertised.

Thank you, LL, for sticking to your word.

Stop being so hysterical and hateful Myra, it's a good thing they decided to release the RC earlier rather than wait ages for the complete rollout.

Also Travis you one of a few people that actually uses the volatile and terrible implementation of XML-RPC. I'd deeply recommend moving to a llHTTPRequest design.

XML-RPC and httpRequest are not equivalent, one can initiate from outside SL and the other one is only from inside to outside SL. So use of XML-RPC is not really a choice or a pleasure for the scripter, it is a necessity depending of the architecture.
I had this issue on the beta gris long times ago, for me it is working in some of the few regions running under 1.24, but maybe this issue was not corrected, I never really had an answer in the Jira about that.

Many things went wrong yesterday and there was no notice to show that LL is aware of the issues.

Builders were reporting refusal of prims to accept textures, reversion of permissions, snap-back of moved prims, and failure of rezzed prims to take up their correct rotations upon rezzing. People who normally do not crash were crashing repeatedly. Ghosting of avatars and prims -- ghosts visible to some residents and not others -- was also happening a lot. I am not just whining about things happening to me, I am reporting what I learned from group IMs and what I observed as an admin in 3 sandboxes in different regions.

This morning --- well before outage time! -- I could not log in until after multiple attempts I was shunted to Help Island Public. There I found dozens of other people who were also unable to get into their home region or desired region.

Please when these things are going on, post something in the grid notices so that we know you are aware of the issue and working on it...not blithely doing something that is making it worse and worse. Or if you need to be doing something that is making it worse, please tell us that too.

Many things went wrong yesterday and there was no notice to show that LL is aware of the issues.

Builders were reporting refusal of prims to accept textures, reversion of permissions, snap-back of moved prims, and failure of rezzed prims to take up their correct rotations upon rezzing. People who normally do not crash were crashing repeatedly. Ghosting of avatars and prims -- ghosts visible to some residents and not others -- was also happening a lot. I am not just whining about things happening to me, I am reporting what I learned from group IMs and what I observed as an admin in 3 sandboxes in different regions.

This morning --- well before outage time! -- I could not log in until after multiple attempts I was shunted to Help Island Public. There I found dozens of other people who were also unable to get into their home region or desired region.

LL, and in this case maybe specifically the rollout team? Please when these things are going on, post something in the grid notices so that we know you are aware of the issue and working on it...not blithely doing something that is making it worse and worse. Or if you need to be doing something that is making it worse, please tell us that too.

Bau -- I am not sure about yesterday's issues, although there *was* an issue with one of our VPN machines and system configuration that was causing strange problems for a subset of regions. They wouldn't have been grid-wide.

Everybody -- the final stage of rollout to 1.24 is currently on hold, waiting for the system problems to recover. (The current outage is unrelated to the 1.24 rollout.)

Myra -- we had a lot of pressure in both directions. Some folks wanted the 1.21RC client released earlier so that they would be able to start playing with Mono without having to use the Preview Grid client.

We knew this change was going to be awkward, which is why we didn't want to release the client until after the server was out. However, with the way the scheduled fell out, we were looking at the RC not being out until next *Tuesday* (what with this being a 3-day weekend) if we didn't release the client before the server rollout was finished.

To put it simply, this is unacceptable. Your partial rollout of 1.24 has broken the grid. Vendors in the new 1.24 regions cannot talk to their servers in 1.23 regions. Consequently, I'm plagued by dozens of failed transactions and angry customers. I will personally tell all of them to IM you if this isn't resolved today.

Allow me gently to suggest that threats are perhaps not the most productive form of engagement....

What do you mean by "talk to"? What communication method are you using?

The 1.24 rollout is now in progress, so this is probably about to be a non-issue. However, if you see problems like this, the right way to proceed is to file an issue on the public JIRA (http://jira.secondlife.com) in addition to noting it in the forum here.

It was not a threat. It's a simple statement of fact. For every customer complaint I received about vendor transactions, I told them the honest truth. The vendors won't work properly until 1.24 is completely rolled out, and if they have an issue with that, talk to you.

It is a widely known problem that llEmail does not work between 1.23 regions and 1.24. This was tested and reproduced many times on the beta grid, and on the live grid yesterday and today. Since email is the only way to communicate between objects in different sims, and every networked vendor on the grid has to do just that, the partial rollout has caused major problems.

Please update us... Thank you!

Prospero posted a few minutes ago saying it is in progress. *grins and bounces*

I wasn't aware of this issue.

I've been digging into my RPC problem through the night. It now appears that its not RPC that's failing at all, but rather llEmail, as the poster above is describing.

I'm also attempting to send an email from an object in a 1.24 region to a object in a 1.23 region - I'm never receiving a reply, which made my script behave like RPC was failing.

BTW the llEmail issue is logged under this JIRA. Probably want to add your comments there

http://jira.secondlife.com/browse/SVC-2918

According to Nyoko Salome's comments, it looks like your vendor problems are only going to get worse as things are rolled out!