-
-
Notifications
You must be signed in to change notification settings - Fork 296
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Update to 1.50.1 ended in 503 Service Unavailable #1598
Comments
I have exactly the same problem and also informed @theCalcaholic |
@cb-studi
|
@saschabrockel |
My NCP instance is old (from Q1 2020 I guess) and the last updates all caused problems for me... But it is also sad that I just learned today that people who use Docker should never use ncp-update which I have done back then and it also was activated for auto-updating... maybe that has caused some problems over time. |
Edit: never mind, missed that you guys are using docker |
The v1.50.1 image is currently removed, can you guys remove the container & image for v1.50.1 ? 🙏 |
@saschabrockel could you please follow the steps described in #1593 (comment) and see if that works around your issue as well? |
@theCalcaholic did not work. It does not change anything. But I'm still not sure if we have the same issue. I guess it is something with the database. That was often a pain in the... |
@saschabrockel Sorry, I forgot to mention: You need to restart the container afterwards |
@theCalcaholic Sorry, and I forgot to mention that I did that of course :D |
@theCalcaholic have you recognized my missing MySQL logs? I had trouble with another MariaDB container for another application and I know it was helpful to add these both as environment variables:
|
It could be, is it related to #1577 perhaps? |
I am so dumb but I guess this restore completely f*cked me. What kind of restore is this??? All files are gone. Of every user. I would have never thought that this could happen and I have no goddamn copy. I have used the ncp-backups folder with a backup of the 01.09.2022 with 900MB. While the restore was running I saw that there was a backup of the update from yesterday... but the goddamn restore also deleted this one and the whole folder I feel I'm done with NCP now... how should I ever trust again if a restore deletes all files of all users. How should I explain this to all users now? And the best thing is it is still error 503 service unavailable. I'm so fcking done @ZendaiOwl couldn't u mention this... #1417 @theCalcaholic any idea on what happened there and how to get the files back? |
@saschabrockel Yes, I'm afraid it wasn't the restore that deleted your data, it was the broken update (compare #1577). I'm not yet entirely sure why the data has been deleted, but I know how it happened for the database. My focus has so far been on preventing this from happening for all the users which didn't update yet. Now I'm investigating ways to recover. This update was really disastrous and should never have passed my tests |
@theCalcaholic I have not looked into the folders but it looked fine from the outside after the update. Also as you have mentioned the I had a backup without data included and I guess I deleted everything with this restore. I am trying to use foremost and multiple recovery tools now to get something back... I thought using a RAID 1 would be enough and would have never thought this could actually happen. Also weird is that it was said that update 1.50 caused the problem but if I have a look at what Watchtower did there it was updated instantly to 1.50.1 and as I have read that should not have deleted the database. |
@saschabrockel I have been mentioning it non-stop .. There are warnings littered over the related topics in the forum and the issues here. I am one person, working full-time and studying full-time, I can't do everything for you |
@ZendaiOwl is a volunteer helping with NCP in his free time. That he is as active and knowledgeable as is the case, is amazing. He's not support staff and can't be expected to guide everyone through fixing their NCP instance (nor know any issue that has been brought up within the project). |
Yeah, I understand it. Warnings in forum okay but who would goddamn think that a restore or update can delete all of my files. My whole university career with files like my master thesis was only saved there. You can imagine what this means to me. I have no idea to rescue and I really need to get this back... I have no copy. I would really really appreciate any help I could get. |
@saschabrockel I'm very sorry to hear this. :( |
Btw, I know it's not of much help to you, but I have created an early draft for a fix for ncp-restore (still requires testing, though). |
@theCalcaholic for Nextcloud I don't have any backup except ncp-update-backups. The folder you mentioned in the other post (the automatic update backup folder) was deleted on the restore. I have luckily some data that is not gone like password files because this would literally cost me everything but my friends had some data there and they were only using it per browser. The same applies to my university stuff of 10 semesters... I mean I have already finished everything but I had tons of information in it and really liked to have them to see things I might forget. Everything else is not a problem for me. It is really just that one big thing. I deleted from the university OneDrive after I left the university because I did not want that they have all data there... I guess that was a dumb decision. You can be sure that I'm going to learn now how to back up NextCloud the best way. I have read 2 or 3 years ago about rsync and stuff but I was just too much because I was a bloody beginner. And until today it was no problem. Yeah but I'm still happy to see the early draft. I wish nobody to have this. Imagine that would happen with a company's data. I know you also do have not this much time and I'm sorry for my rage on some posts before. It was not against @ZendaiOwl. I just could not believe that a ticket of January which I have seen too late cost me this... I guess I need to set up the whole NextCloud instance from scratch because the restore didn't even work and there is literally nothing left I guess... I'm trying to use ext4magic now to get some data back but only strange things and little files are found. Tons of .gitkeep, .txt, .PHP and .unknown files. I have the feeling that the files were deleted and overwritten by the new empty folder for each user... |
I'm afraid there's only little chance of getting your data back :/ I don't mean to excuse the bugs you were encountering, but I can't overstate the importance of a reliable backup strategy. As a starting point, I'd recommend mount another host directory to the container, e.g. That's only a starting point, because to have really reliable backups, they should at least reside on a separate disk, if not a separate machine/location. But it's a good start :) |
My problem is or was that I do not have so much free space. I thought having this data on RAID 1 would be enough, to be honest, and I would never imagine that this could happen. And that is the point where I don't know how to have a small or good backup because I can't have 1 TB of data on three separate disks and I guess you understand the point. But I guess you would say I need at least one drive more for my data saved on RAID 1. The point where you could me a bit is that I don't want to attach a USB hard disk all the time because it is not good for the lifespan of the hard disk. The other disks are NAS disks so it is no problem. But how to backup for example every day with a USB hard disk but only turn it on at this time for this purpose? And I am not sure yet if I might not get the data back. ext4magic is running now for 20 hours I guess and brought me 32GB back but I don't know exactly what it is yet. Many files seem to be so old I cannot imagine that they were brought back. But the problem is that I have not that much free space on the hard drive the recovered files are now written to and I won't be able to pause the process and say use now another hard drive to save it... But it seems to be a good tool for Linux users that might have the same case as me now. If anybody reads this here they should give it a try or ask me if they need a short instruction. You can mention my comment here if somebody else will tell you about a similar case. |
@saschabrockel I wouldn't save on the backup disk, that's too important. Backups are more important than your raid setup so if you already have a raid 1, consider migrating to a single data disk and a single disk for backups. One option for backups that's very disk space efficient are btrfs snapshots.
These snapshots can then be copied to a second disk (which is better than raid, because it protects against accidental deletion) |
Watchtower automatically updated the image in the night and as soon as the update was over I got multiple notifications that NCP is down with error code 503. When I go on the website I only see the following:
It looks like there is something wrong with the database or its logs.
Edit: there are no log files anymore in the database folder.
System information
NextCloudPi diagnostics
Nextcloud configuration
HTTPd logs
Database logs
Nextcloud logs
The text was updated successfully, but these errors were encountered: