Loss of State affecting some users / some SmartApps (was Death knell for RM today?)

Appologies for being thick but where exactly is this.
I’ve rummaged around the app but cannot seem to find what you are talking about.
I did have a piston playing silly beggars so I have rebuilt them all.
Thanks for this new feature. Hell of a time saver.

Do you guys work for a living because your response to issues on the platform or feature requests is unbelievably quick.

Thanks again.

EDIT. I think I have just worked out what you mean.
It’s the arrow pointing to the right?
Doh.

3 Likes

If I am seeing this correctly, you are looking at the Smart Apps menu in the Settings. You need to go to the Marketplace > My Apps to install it… Please let me know if that worked.

TO THE STAFF OF SMARTTHINGS;

Just received word that the all the connectivity issues have been resolved. Thank you for that.

I am looking for an update as to IF or WHEN our lost data will be restored. Any probability in the next 24 hours?

The last time this happened we were promised that our data would be restored but time drug on and nothing ever came out of it. Some of us have spend untold hours building our home automation systems and loosing our work (for the second time) is a difficult thing to understand. Is the problem REALLY FIXED this time?

7 Likes

Is it not possible for us to somehow develop a way of backing up all our user data locally? This has always been my biggest fear with this system. Other than that I love it! I understand growing pains happen. I’m not about to add fuel to that fire. Giving users the peace of mind of some sort of backup would be awesome. I don’t know where to turn for this. Is it a smart app developed by Smartthings or the community? Thoughts anyone?

1 Like

Odd, that’s what I did but I tried again this morning and it worked! Maybe it was a cached list or something. Oh well, I guess this was the kick in the rear I needed to move over to CoRE. I love the if/then approach over rules and triggers anyway.

thanks

Not at present. It has been requested multiple times. for example, last year we were promised a “migration tool” that would simplify an account moving from a V1 hub to a v2 hub, but that has never been delivered.

1 Like

To clarify, there were two outstanding incidents this morning. The issues accessing Graph were resolved. We are still investigating the underlying cause and scale of the database change and reversion yesterday. Separate (more detailed) community facing post-mortems should follow once we have this wrangled.

3 Likes

so does this mean we need to rewrite our broken rules? a bunch of mine are firing off whenever they want to because they are headless.

If you are using CoRE, the latest update appears to restore pistons (thanks @awesomebadass @ady624!). I am not yet sure about other SmartApps (but should have updates soon-ish)

7 Likes

What this means is. …

Rule Machine is not a supported app. Not by ST and certainly not by the developer. He left us all high and dry on that one.

I’m actually amazed that RM has lasted this long. I’ve been saying four months that RM is a dying animal and that eventually an update would be a final shot.

Yesterday was a warning shot. RM is going to be completely dead soon.

And in my opinion, if after tomorrow, if you continue to use Rule Machine then you are choosing the problems that are coming for you.

I’m not being an ass. I’m being realistic. I’m not saying use CoRE. But yesterday was a notice that you had better find an alternative.

Good luck.

Edit

  • this is not meant to any one person, but to every user that has RM controlling something in their system.
12 Likes

Agreed. For any app that’s no longer supported it’s when, not if, it will break IMO. I understand the idea of “if it ain’t broke don’t fix it” but I’m a little surprised at how much animosity there is from people who have known that RM is dying for quite awhile.

Look at how quick @ady624 was able to come up with a fix for CoRE…turns a cliff into more of a speed bump for users with a supported app. My opinion anyway, and I know what they say about opinions, especially when I’m sitting here not experiencing any of the issues that everyone is complaining about :fearful:

3 Likes

@Aaron, perhaps the fact that RM was running actions left and right after having lost the conditions (state loss) explains the elevated number of API requests? Just a thought…

8 Likes

I agree. When it comes to third party smartapps, if the developer is not at least somewhat active in this forum, I don’t touch it.

Kind of like buying a repossessed mobile home, sight unseen, in Mississippi… You may be able to live in it… But is not going to be a fun time.

1 Like

This motivated me to convert all of my rules in RM to CoRE since a good number of my RM rules were headless. Not only is @ady624’s awesome app able to recover broken rules, but the screenshots from the Core Dashboard are an awesome idea. I took screenshots of each of my pistons and uploaded them to my Google Drive for the next “state” failure.

2 Likes

Ya know, I still use RM for 11 things. All could be moved to CoRE of course. But they have been 100% solid. Yesterday 5 of my 11 either went headless or lost portions of themselves. I will be recreating those in CoRE and then recreating the other 6 in a disabled state in CoRE.

The simplicity (relatively) of RM keeps me from transitioning all rules. Only if I can improve on them with CoRE have I migrated them. I don’t blame RM or Bruce for any of this, he had a great app, great support, and a great reason to bail. This is all on ST.

@ady624 got the restore function rockin’ but honestly, that doesn’t prevent the potential for all loss. I’m still curious how it works btw :wink: but in reality I only feel slightly safer having rules in CoRE.

What we really need is a “backup account” option in the IDE, I can’t imagine it would be that hard to implement, just save all the devices and associated info, the Smart Apps and dht’s, and their state info. It’s all just stored in a database after all. A restore option would backup existing, purge all account records, and dump in the backup as if it’s a new account. I don’t care if it can be backed up local, in fact that’s a bad idea, too easy to introduce “badies” and cause other issues unless it’s encrypted, just back to ST infrastructure, that’s hopefully separate from the noSQL AWS stuff.

I guess what I’m saying is I use RM still because I don’t like all my eggs in one basket, especially considering the issues we’ve had with ST over the last year +.

1 Like

This with bells on! What has also become apparent from this thread is how many people are still using rule machine and I get it, Rule Machine was awesome… but software not actively being maintained is always going to be doomed no matter how great it was at the point it stopped being maintained.

It’s taken me a little getting used to CoRE but once you do, you’ll see it’s every bit just as powerful, if not more.

To me there are 2 major issues here:

  1. The obvious state failures that caused data loss.

  2. The fact that it’s over 24 hours later and there’s been no indication they can recover our data. Jobs would be lost if that happened in my company. The second they noticed there was data corruption, the db should have been locked and a restore should have happened. Yes, some new people or data who made changes might have been impacted, but it would have been far less impactful. Now, it’s too late. I already removed Rule Machine from my setup and converted all of the rules to CoRE. How are they going to possibly restore data with so much time passed? It would now cause even more of a problem.

7 Likes

Exactly, that’s why per user backups are important. The sledgehammer of freezing the entire db and restoring is not conducive to this kind of environment. Need the surgeon for this one. I’ve spent enough time in SQL and it’s derivatives to know this must be possible with the noSQl derivatives as well. Right? Anyone? I know we have some DBA’s in here… :slight_smile:

This bugs me:

Wikipedia:
Apache Cassandra is an open source distributed database management system designed to handle large amounts of data across many commodity servers, providing high availability with no single point of failure.

What’s ST doing wrong?

.

2 Likes

I think we are all just guessing at this point, but I would add that if you introduce a terribly jacked schema update that blows things, or other bad code that interacts with the db…doesn’t matter about high availability. It will make that mistake highly available eventually. The scope of this failure is larger than the status page saying its a subset of users.

2 Likes

Well, they could always use just about every db in existence that has multi user interface… Each account is able to be restored.

Here is a very simple example… Microsoft exchange server operating within a NT environment.

2 Likes