Duplicati, Cloudflare API, Drive Crash, Peewee

  • Yet another problem with Duplicati, it seems that it simply hangs with some trivially ridiculously small tasks. Like in this case downloading file (8.43 KB) and that's it. It basically halts, nothing happens, ever. Only way forward is to kill the process. And I mean, after 24 hours, not like after 5 seconds. I know how ridiculous claims people make, so that's exactly why I logged in with parallel session and downloaded ALL backup data, while waiting for that task to finish. So what ever blah blah, you're trying to lie / claim, that it's not true, well your claims are invalid. Log says: "[Information-Duplicati.Library.Main.BasicResults-BackendEvent]: Backend event: Get - Started: duplicati-..." and then it hangs. When I write crappy code, I'm proud of it, because so does everyone else. Haha. - I guess it's back to good old method for a while. Create one 7zip packet and download it nightly. Sure, bit of overhead but who cares as long as it isn't several terabytes. Isn't that exactly the reason people perefer unlimited data plans? - This is an issue which has been resolved a long time ago, but was really annoying at the time.
  • Cloudflare API, interesting. It was broken for several hours and returned 500 Internal Server Error, but now it's working perfectly. I didn't change a thing. Modified my script to use extended retrying every 60 seconds. It's pointless to retry too often.
HTTP Error 500: Internal Server Error Date: Thu, 26 Apr 2018 14:18:15 GMT
Content-Type: application/json; charset=UTF-8
Content-Length: 158
Connection: close
Set-Cookie: __cfduid=d8319aef7fc34787b0972d576309c8f6f1524752290; expires=Fri, 26-Apr-19 14:18:10 GMT; path=/; domain=.cloudflare.com; HttpOnly
Strict-Transport-Security: max-age=15780000; includeSubDomains
Served-In-Seconds: 4.701
Expect-CT: max-age=604800, report-uri="https://report-uri.cloudflare.com/cdn-cgi/beacon/expect-ct"
Server: cloudflare
CF-RAY: 4119ad588a1068de-CDG
  • Even more problems with Duplicati, now it created broken archive (why, that's also a great question!), and crashes on it every time. This means that manual fixing operation is required, which absolutely sucks. And after that purge-borken-files, you'll need to rerun repair and let's see what next. 100% hang, not being able to repair. Only way forward is deleting whole backup set and restarting. That really absolutely and totally sucks. Automatic software should work automatically. I don't even begin to mention how much fail it is, if automated software sucks totally. Backup says run repair, repair says run purge, purge says nothing wrong. Seems that the legendary state machine is broken, yet again. - It's fun to enjoy every day, bad or extremely bad software. - Deleted whole backup set and history, now backups are working again. But let's hope it won't break again. - And about negative things, one version had some kind of invalid references to libraries and refused to work at all. And one version had help feature which crashed it with many of the help requests and so on. - Also the scanning local files status reports are now broken. It shows that 2 files and (0 bytes), and this information won't get updated at all as it used to be. - There are still a few kinks with Duplicati, but in general latest versions are working a lot better. Of course this is normal in software development. See Duplicati Forum for latest discussions.
  • Took the 3TB crashed backup storage drive a part. It's literally crashed drive, many of the disk platters are totally scratched, so no wonder it didn't work anymore. Still wondering if external shock could have caused it? Head crash should be extremely rare with modern drives without getting shocked.
  • Everything is so broken in this life. I just updated Peewee ORM with latest version and FTS indexing is broken, can't even create FTSModel with SQLite3 based models at all, as well as local threading (threadlocals) isn't working anymore. Well well, had to roll back. I don't have time to get stuck with stuff like that for days when things just should be working. That's just the problem when getting things done. It's different situation, if you're looking for trouble and have reserved a few weeks for wondering something. Instead of getting it done in very limited time. - That's actually why I like KISS solutions, the less there are things to get broken, the better everything will be working. Anything cool, complex, extra dependencies, really bad idea. Maybe there's documentation about the changes, but right now I don't have time to read it. Well, I'll try to get it done some evening. Hmm, yeah, it seems that there are major differences between 2.10.1 and 3.3.1. So it's just a good idea to make this independent process. Something which just can't be "updated", but requires proper migration.

2019-09-15