Enonic version: 6.13.1, but this also happens with 6.15.4 OS: Manjaro 18.0.0, but Enonic XP is running in a Docker Container
I’m migrating an existing Enonic XP site to Docker. The current maintainers of the site has made a system dump using the Data Toolbox application. I have spun up a fresh instance of Enonic XP 6.13.1, the same version as the system the data was dumped from. I logged in as Super User, installed Data Toolbox, uploaded the system dump and loaded it. It spun for about 2 seconds before a box popped up with “500 Server Error”.
The log is showing this error:
m.enonic.xp.repository.BranchNotFoundException: Branch with id [draft] not found
The site appears to have been loaded correctly in spite of the error, and I don’t see anything missing. I’m not deeply familiar with this website however, so I could be wrong.
Why does this error occur? Is it safe to ignore? Is there anything I can do to prevent it?
It is not safe to ignore no. The system dump was not loaded properly.
First, can you confirm the steps/versions under:
You have a source installation with
Enonic XP 6.13.1 ?
Data Toolbox 2.2.3 ?
You do a system dump
You load this system dump into a target installation with:
Enonic XP 6.13.1 ?
Data Toolbox 2.2.3 ?
Could I have access to the dump in question?
If not can you look at the dump.json at the root of the dump folder (check that it is present) and send me a message with the content of the file?
Because from the logs it seems that the dump is incomplete/missing files. Something must have happened during the dump or the dump was downloaded while still being in progress. A quick solution to your problem (and it would confirm that it is linked to this dump in particular) would be to generate a new dump and load it.
I can confirm that the Enonic XP versions are both 6.13.1 and that the Data Toolbox version of the importing instance is 2.2.3. I have requested the version of the exporting Data Toolbox instance, but I haven’t received it yet.
We have generated 3 separate dumps over the past 2 weeks and they all trigger the same error.
I unfortunately can’t share the dump, but here’s an overview of the contents, with binary and node folders pruned for brevity: https://pastebin.com/raw/f6HYBZQU
The exporting Data Toolbox version was 2.2.1. I have downloaded and installed version 2.2.1 to a fresh local containerized Enonic XP 6.13.1 instance, and it’s still crashing with 500 Server Error.
We can see that your dump has finished successfully but it is seems to be missing at least one file: /node/bc/c0/e1/bcc0e1e7d0466ff8ef73f04f0d5c8d5de88a959f
I manage to reproduce your error if I manually remove a similar file from a dump. But I don’t manage to generate a dump with a missing file. Even if I remove manually the blob in my $XP_HOME/repo/blob/node/, restart the server and do a dump: the dump simply fails but yours has finished.
Is there anything in the source server logs when doing the dump?
Is there anything in the source server logs when retrieving the dump?
Is the following file present on the source server: $XP_HOME/repo/blob/node/bc/c0/e1/bcc0e1e7d0466ff8ef73f04f0d5c8d5de88a959f (Assuming that your blob folder is located at $XP_HOME/repo/blob)?
I assumed it was the case but just for confirmation, check that the following file is missing in your dump: /node/bc/c0/e1/bcc0e1e7d0466ff8ef73f04f0d5c8d5de88a959f
And lastly, is the file present in the dump on the source server: $XP_HOME/data/dump/dump-2018-09-25T19-33-35/node/bc/c0/e1/bcc0e1e7d0466ff8ef73f04f0d5c8d5de88a959f
Unfortunately I have no access to the source server.
After deleting a blob, did you dump using Data Toolbox or toolbox.sh? Our dump was generated using the former. Perhaps it’s happy to ignore the missing blob?
I have now switched to importing using toolbox.sh rather than the Data Toolbox application, and now the import succeeds 100% of the time (no crashes), but the result stats shows the missing file you’re referring to:
If this just means that 1/13987th of the cms-repo data is missing, we’re willing to ignore it. Is there any way this can cause problems later on? Mysterious crashes, hangs, etc.?
If this is too dangerous to ignore I suppose we’ll have to ask for a new dump.
Yes, it is not critical.
Because from what you sent it is an old version (not the current used version) of the content “/website”.
The only visible impact for you will be that old versions wont be displayed in the widget “Content Studio”/“Version history” and you will not be able to rollback to this old version.
But to my knowledge, no, there will not be any problems linked to this