Recently I wrote a lot of book reviews on Logseq, and several days ago I found that one of my pages only showed half of the content. I checked the page history and did recover it from that. and today this happened again. I did nothing which I think will cause these to happen, I didn’t install any plugins or open it at iphone and mac at the same time at these days. I always edit my content at MacBook to ensure the stablility. But now logseq seems not stable at all.
Yes, we see this on the Windows desktop version as well. Most often it seems to be caused by some random occurrence - and things seem to disappear from the graph. Sometimes re-index will fix it. Often you must force stop Logseq (with Task Manager in Windows) and restart. Most of the time the underlying MD files persist - although the most recent two occurrences resulted in “partial data” in an MD file and in the 2nd case it seems Logseq “removed” an MD file, perhaps because the “glitch” made it think it was no longer necessary. Since the software is a Beta product - such issues are somewhat to be expected. But, as you stated, we have decided Logseq has enough inconsistencies not to put “production data” into it. Hopefully things will get sorted out by the time v1.0 arrives - then we will re-evaluate the product as it shows a lot of potential.
Thank you for your explanation.
From your explanation it seems that there’s no quick way to avoid such things. I really enjoy the note-taking process within Logseq. Will keep on track of the new version, but not gonna use it lately cause most of my writing are quite important to me.
Same experience, I tried to move Logseq Data from one repository to another, results in a total loss of data, pages still available but database obviously totally corrupted. Product in this state is not recommended for professional use.
I too have experienced data loss but couldn’t pinpoint it. I quite often notice that link refs become errors as whey they point to has gone missing. I have considered putting git under logseq to track file changes but haven’t done so yet.
Do you use any plugins? Some plugins have issues with this.
Had same experiments before. Then I write an script to auto commit on github every 5 mins.
No plugins are installed.
I look at the files with vscode every time a weird glitch happens and never actually lost data. Would not use it without, it’s also a great history view!
TL;DR - I trust Logseq for what it is: a beta app, and I’m comfortable with this. And yes: I might be gone OT, sorry
I haven’t experienced data loss, not yet at least… but reading version number of Logseq (0.6.7 as I’m writing) makes me think that we’re in some kind of “early beta”. If you think of Logseq as a beta software, you’ll be amazed by its quality and stability (and community). But if you think of it as a productivity (stable) app, then you’ll probably be deceived.
A couple of years ago Roam “created” a new type of tool for thought and many devs dived into this new “market”. That’s why there are many “alpha” (see AnyType or Athens, just to mention apps I ues(d) ) or “beta” (see Obsidian, v. 0.14.7 as I’m writing) apps that have thousands of users. Quite a strange thing in software market, but there it is.
The problem, as far as I can see, is that people who need and rely on this kind of software use it for professional purposes (academics in my case) and any “problem” you may have has a great impact on you (think about loosing a page that costed you months of research, maybe traveling to reach a document…).
Just my opinion/experience, of course.
Yeah, agree. It’s my misunderstanding about Logseq, I thought it was a stable product according to it’s mature UI and neat interaction design of the Mac version or iOS version of the app. But if it’s not stable until now, what are you recommend to use it with?
I, personally, would recommend using it only with “test” data. At least that what we have decided. There are too many inconsistencies and too much missing core functionality to consider investing time and effort (and risk) using it with any “essential” data. And of course, as a Beta product - that makes perfect sense.
We encourage students to try it - to use if for basic note taking and basic task management. It is quite an eye opening experience for most of them. They find that simply trying to learn all of the concepts about linking, tagging, backlinks, unlinked mentions, etc. and how to use them requires a huge investment of time.
They don’t use Logseq as any kind of permanent “knowledge management” platform - but it does make them THINK about the complexity of the process. It opens their eyes with regard to how much time, effort and planning it would take to design, build and maintain their own “personal knowledge base” in the future - regardless of what the “app of the day” happens to be at that point. And thinking / learning is always a good thing.
The good point of Logseq is that is stores everything you write in markdown directly into your hard disk. This means that if you have some kind of backup (at least) of the folder(s) you leave your work in, you should be fine.
I use Logseq as a “writing machine” on steroids, meaning that all my original research material(s) (i.e. books, papers, data and so on) is also in other folders in my hard disk.
Should Logseq “eat” my files or part of them, I would “only” loose my editorial work and not every single source I collected.
Hope I got the point
Totally agree with you
This is an unfortunate experience. I have so far only lost data through my own stupidity (git reset). If you want to continue to use Logseq, you might be interested in a solution like Syncthing, which you can configure to version your files (as a form of back-up) and sync them to other devices.