Update recently? deleting

While original MongoDB is here https://github.com/mongodb/mongo

Percona website also has .deb packages etc

for installing

It is also possible to export all MongoDB data to JSON files https://github.com/wekan/wekan/wiki/Export-from-Wekan-Sandstorm-grain-.zip-file

For corruption in general, solution is to do daily backups. Many databases like MySQL, PostgreSQL and others can have database corruption also.

at least daily. More often, if possible.

Well, some years ago when I worked as SysAdmin/Cloud Architect, I helped coworker, that had Mac laptop. Upgrading to newer version of macOS had corrupted all local MySQL databases. I could not recover anything from those, only solution was to download newer database dumps and reinstall all local databases that where used from Drupal development.

I understand. And thank you again for your thoughts and all the detail. Really!! For the record this is AWS EBS so its not data corruption or anything related. I set aside additional ebs drive just to host the mongo data component not use the root drive. Now You mention several things in the thread before your last post. First you mentioned that the software was trying to remove a collection and you were surprised. You also mentioned or asked actually if this system was exposed to the internet. Since the issue was database related and since Mongodb has a history of strong arm warning folks not to every run mongo without auth (yes they get ugly) then this was a dialogue I wanted to explore. Yes indeed local host access is very secure. When I started this thread I asked about a code change that got released recently as I had three system with related symptoms (logical conclusion) right? ok well if that was not it and you saw the error about issues that did not make sense about removing a collection (your reaction) then if it not a mongo internet data breach then that kind of leaves two path outside of that. One, that the wekan software corrupts after a certain point (which as we have high traffic (dev) and low traffic (demo) and this happened all at the same time that does not makes sense either. Logs show nothing related to a denial of service issue. In point of fact one of these systems is on secure gov cloud even. Which going further down the rabbit hole could indicate that someone hacked all three systems (hacked in this case means a virus or other on wekan. Occam's razor tells me this is not likely, so I was just looking for the simple reason. I get backups. Of course. If my data is transitory (puting aside login accounts and access tokens obviously) then I can and will replace all the data anyway automatically in a second. Why do you think then why am I backing up? What in your opinion broke and why?

Do you have some script, that by running it you can duplicate Wekan breaking?

I could not figure out exact reason from above info. It's just guessing.

It can be Wekan, Meteor or MongoDB bug, or bug in some dependency code

And for any bug, it needs minimal code example to reproduce bug

For example, someone recently reported that exporting one board did not work, when using Wekan Snap. Exporting other boards did work. He sent copy of that database to me, so I could test it. I noticed that with Wekan Docker version, it was possible to export that one board. The fix to Snap was to upgrade to newer version of Meteor, and then publish automatic Wekan Snap update.

I have no idea how exporting one board and version of Meteor is related to each other.

So in general, I care only about those bugs that can be reproduced.

Other bugs could come and go depending on status of each dependency.

I do trust worldwide community of all Wekan dependencies, that everything will improve. And if some dependency package does not work, I can switch it to some other dependency.

thanks. No I do not have some way of making it break (yet).

Question: if it was something to do with meteor, could I even fix it? Meaning would I be able to upgrade Meteor inside snap myself? I thought that the version of wekan was tied to a particular version at any given time of meteor and in turn Node.js? If I could upgrade would not the next automatic upgrade of wekan either overwrite that or get stuck if the version of node or meteor was newer? My issue is if I could figure out this does not mean that I could fix this long term, or are my assumptions wrong with that?

ok well the issue came back again

refreshing the gui in the browser and I see all the archived boards I deleted

where do I look for the actual errors again?

sudo snap logs wekan.wekan 2020-02-28T22:54:12Z wekan.wekan[30813]: > Starting add-description-title-allowed migration. 2020-02-28T22:54:12Z wekan.wekan[30813]: > Finishing add-description-title-allowed migration. 2020-02-28T22:54:12Z wekan.wekan[30813]: > Starting add-description-text-allowed migration. 2020-02-28T22:54:12Z wekan.wekan[30813]: > Finishing add-description-text-allowed migration. 2020-02-28T22:54:12Z wekan.wekan[30813]: Meteor APM: completed instrumenting the app 2020-02-28T22:54:15Z wekan.wekan[30813]: (node:365) [DEP0005] DeprecationWarning: Buffer() is deprecated due to security and usability issues. Please use the Buffer.alloc(), Buffer.allocUnsafe(), or Buffer.from() methods instead. 2020-02-28T22:54:43Z wekan.wekan[30813]: Exception while invoking method 'ATCreateUserServer' Error: getaddrinfo ENOTFOUND mailserver.examples.com 2020-02-28T22:54:43Z wekan.wekan[30813]: at GetAddrInfoReqWrap.onlookup [as oncomplete] (dns.js:64:26) 2020-02-28T22:56:15Z wekan.wekan[30813]: Exception while invoking method 'ATCreateUserServer' Error: getaddrinfo ENOTFOUND mailserver.examples.com 2020-02-28T22:56:15Z wekan.wekan[30813]: at GetAddrInfoReqWrap.onlookup [as oncomplete] (dns.js:64:26)

at this point the gui is not stateful with the mongodb

do not see anything in the mongo logs