XML-RPC enabled scripts need a reset with 1.7?

Hello all,

I can't trust my own scripts to work properly all the time, so I'm always suspicious about my own work, but I found it strange that all my XML-RPC-enabled thingies started behaving badly since 1.7 (the only thing I couldn't test in the Preview grid!) and needed a reset.

Now I'm aware that Ginko, SL Boutique and SL Exchange, among others, also ask their customers to reset their scripted boxes/ATMs. And the Greeter "in-world request dispatcher" (which I always suspected to be a simple device that IMs an available avatar and nothing fancy...) seems to have stopped working as well.

Since the Hotline is crammed full with questions, and the Lindens will need to work overtime to deal with all of them, I'm very reluctant to post my question there So, does anyone confirm that all RPC channels have changed and definitely need a reset? (in some of my devices, to make sure they really worked, I even recompiled them, just to make sure). Apotheus at SL Exchange is not so categorical and only recommends a reset.

Thanks for any input you may have!

Well I can't 100% confirm it, but I'm pretty sure they all need to be reset now .

I am having the same problem. Unfortunately for MetaAdverse, it means getting 100's of publishers to reset their signs manually.

Yea...I've also noticed that my llVolumeDetect scripts needed reset.

Is there some global way all of these types of scripts could be reset, or is there risk to upsetting existing content because of it? Sounds like the stall was uniform though.

It would be nice, Torley, but I don't think so

I just talked to Rathe too and it doesn't look that way. Ugh... I can imagine how frustrating it'll be, especially for people with a lot of stations.

Not directly related, but I've also found all my volume detect objects have needed a reset too. The collisions still worked, but they weren't phantom any more. Fortunately all of them I have could easily be reset, but you might like to check.

I've noticed several of my particle generators also needed a reset... not sure if there is a connection there.

It's kind of worrying that so much has needed to be reset after this latest patch. It doesn't scale at all well for the future, especially the content creators and service providers, if this becomes a common effect of an update.

Objects didn't retain thier keys during the upgrade. No biggie, you just need to get the new keys. In most cases that means resetting the objects so they send email to the server and update thier key, since RPC channel rely on the objects key.

Erm, I don't think that's true. Otherwise nobody would be able to purchase my bubblegum and they are. The bubblegum machines all use an object key to send email to the main delivery object on my land, and they're working fine.

Objects changing keys during upgrades is a biggy, if that's happening. Again, that's not scaleable for the content creators. Good lord, I'm glad that didn't happen with my bubblegum delivery object, otherwise I'd have had to email 2000+ bubblegum machines with the new key, and that probably wouldn't of gone smoothly as LL would of thought it was spam and blocked the emails (they've done it before with slboutique.com I believe)!

I can confirm as of right now that I can NOT send an email from an SL object to an off world address. Nor can I connect from DreamHost to an XML channel that I just opened. Been this way going on 5 days now I believe.

To any host, not just Dreamhost? I'm getting emails from my objects in SL, but I'm not hosting at Dreamhost. I'm suprised the Dreamhost issue, which ever end is causing it, is still going on this many days later.

Yeah the email is a DreamHost address. I need to sit down and test my yahoo address or something. I'm just glad I'm not doing commerce. This is just my in-world news system SLToday.

heh. I guess I stand corrected. I didn't have first hand info, I was just helping a friend debug something through IM and that's what it seemed like happen. Please disregard my ignorant post above

I write and maintain the software for Ginko Financial, and I -can- confirm that XMLRPC keys were lost. That is, the KEYS were not LOST, the connections just need to be remade using llOpenRemoteDataChannel(), much like you would do if the object passed a sim boundry.

The keys themselves, the object's key, etc, have NOT changed. It's just that the XMLRPC connections have been invalidated.

That said, I can also confirm that email OUT of SL and XMLRPC were both offline and/or failing intermittently since 1.7 came out, though it now seems to be at least 98% stable as of today. I was told at one point by Andrew Linden that the XMLRPC interface was completely offline (and I confirmed this with my own code) because "someone was working on it", and this counted for a few hours of the XMLRPC downtime.

This is just what I've learned on the issue from trying to keep my ATMs online over the last couple days. Hope it helps someone.

EDIT: This had no effect whatsoever on outgoing email. Outgoing email service was interrupted for a while, though I have had no long-term problems with it. I do not use incoming email in any way, so I can't say anything about it's current stability, but since object keys have not changed and incoming email doesn't use an event-driven system, I don't imagine any problems with it.

-Hinoserm

My vehicle vendors all keep the server key, and they send an email in-world to that server for delivery. This still works fine post 1.7 with no intervention on my part.

As well, my server sends email to my real-world email to report a sale, and I can confirm I have gotten these emails as well.

Note that I'm not using XML-RPC, and I don't use Dreamhost, whatever that is.

- Jon

Email to DreamHost is definitely not working. Email to yahoo.com is fine. Nor can I connect to an XML channel (recently established) from DreamHost.

Same story here. SL->Dreamhost via email failed. Dreamhost->SL RPC failed. SL->OtherSite email worked. SL->OtherSite RPC worked. OtherSite->Dreamhost worked. Dreamhost->OtherSite RPC worked. So I wonder who borked our connection, Dreamhost or LL?

Something is completely borked this.. I tried to lynx to secondlife.com and it did not work.

Thanks for all your input Yes, it looks like outbout email was just shortly disturbed (peaks?) but is working well. Looking at my logs, I'd say that all objects kept their XML-RPC channel keys, but they were "inoperative", and the scripts needed a reset (sometimes recompile) to get them "active" again (ie. a new llOpenRemoteDataChannel()). Since then, I guess that everything works well again — one of the few areas in SL where I didn't saw a reduction in performance!

You should not rely on those keys... have your scripts use a new key every time they need to talk, or periodically change it. I did this sort of thing for Stagecoach Island. I even have a naming system split over a few MySQL tables, mapping product names and prices to assets and RPC keys.
To reset a huge amount of scripts simultaneously zoom out, drag-select, and find Reset Scripts in Selection somewhere in the menus.