dgtlmoon
Synlogy DS220j - Snapshots from different watches are getting mixed up
Describe the bug Different snapshots are sometimes written to a different "watch", or even get "merged" into a different watch for a different domain, creating baffling results. This (seems to) coincide with one of the websites suddenly reporting an empty snapshot. It does not always affect the same pages.
It happens often (once every other day across ~100 websites being checked once a day).
Version v0.48.01
How did you install?
Docker
To Reproduce
Not happening reliably, retrying use Steps to reproduce the behavior:
- Create two separate watches
- Wait for a while
- Get notifications (via email) and look at confusing diffs.
Example watches:
- https://changedetection.io/share/i5RHrPOFUAga
- https://changedetection.io/share/cE_gl8FkTBwa (This specific pair of watches hasn't had the issue before.)
Expected behavior I expect the snapshots to stay distinct
Screenshots n/a
Desktop (please complete the following information):
- OS: macOS
- Browser: Orion (Webkit-based)
- Version: 0.99.130.2-beta
Smartphone (please complete the following information):
- Device: iPhone 15 Pro Max
- OS: 18.2
- Browser: Safari
- Version: 18.2
Additional context The app is running on a very underpowered Synology (500 MB memory), not sure that has an impact. I'm watching ~100 pages across a dozen domains. As far as I have observed, the problem occurs across different domains. It does not impact the same pages or domains every time.
It's the DS220j: https://global.synologydownload.com/download/Document/Hardware/DataSheet/DiskStation/20-year/DS220j/enu/Synology_DS220j_Data_Sheet_enu.pdf
CPU: Realtek RTD1296 4-core 1.4 GHz Memory: 512 MB DDR4
It's not running anything aside from the base synology NAS features and changedetection.io, so the Synology shows only about half of the memory used, with CPU cycles to spare - but it's excruciatingly slow.
hmm the usage of chrome browsers is very very very tough on a machine if it doesnt have much grunt
I'm not using Chrome, just the plain crawler - the Synology would crash completely if I was running Chrome.
Memory usage isn't crazy - about 50% is utilised, and it doesn't swap (or just a few KB now and then). Caveat: I only have a few hours of data as I had to reboot it after trying to batch-rerun a few watches yesterday evening, which seemed to have taken down the changedetection.io Docker container - it won't try to crawl until later today.)
Memory usage isn't crazy - about 50% is utilised, and it doesn't swap (or just a few KB now and then). Caveat: I only have a few hours of data as I had to reboot it after trying to batch-rerun a few watches yesterday evening, which seemed to have taken down the changedetection.io Docker container - it won't try to crawl until later today.)
any chance you can add some logs in .zip format here?
or better if its possible - to add your backup 'zip' file here if theres nothing important in your data
I sent the backup zip file to your email address. The same issue happened again today btw - first time since last week, I think.
@sebmos were you able to solve it? i think i have somewhat related issue, 1st check/snapshot (after edit watch > stats > clear history) is always correct, 2nd+ check/snapshot somehow getting mixed up and don't show all items (https://github.com/dgtlmoon/changedetection.io/issues/2908#issuecomment-2613346482)
@sebmos, following, cause today (for first time), i got mixture of snapshots from 2 different watches (interestinly they were on same webiste just different pages)
and i found reason for my original bug - empty url in notification list https://github.com/dgtlmoon/changedetection.io/issues/2908#issuecomment-2613993214
so there might be possibility that these two (mixture of AND incorrect generation of snapshots) are connected (say if notification fails for some reason).