Skip to content
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

Gaps in history after a federation outage, fixed by refreshing cache #5915

Closed
ara4n opened this issue Jan 3, 2018 · 2 comments
Closed

Gaps in history after a federation outage, fixed by refreshing cache #5915

ara4n opened this issue Jan 3, 2018 · 2 comments
Labels
A-Timeline O-Uncommon Most users are unlikely to come across this or unexpected workflow P1 S-Major Severely degrades major functionality or product features, with no satisfactory workaround T-Defect

Comments

@ara4n
Copy link
Member

ara4n commented Jan 3, 2018

@erdnaxeli says:

hi, I got a weird case

my HS got disconnected from the network for a week, and on reconnection an user scrolled on a federated room (actually #irc:matrix.org)

but the history was missing two days of messages

I told him to empty his cache and it works better after that

it does not really surprise me, but my users find this behavior a bit strange

and I'm not really sure of what actually happened

to detail: the first user was missing 4 days of history, I joined the room with a user on the same HS, I had only 2 days of history missing

we both emptied our cache, and it was ok

so I thinking something like riot got confused, ask for some history, then stop ; then I went to the room, ask for a little more, then stop too

with an empty cache riot correctly asked for the whole history, forcing the HS to ask to others too

@lampholder lampholder added T-Defect P1 S-Major Severely degrades major functionality or product features, with no satisfactory workaround A-Timeline ui/ux labels Jan 11, 2018
@lampholder
Copy link
Member

Anything that undermines trust in the timeline's completeness is very bad :(

@MadLittleMods
Copy link
Contributor

MadLittleMods commented Aug 25, 2022

This could be a client problem but can also happen easily if Synapse returns /messages with gaps in it. MSC3871 will help with this sort of thing so Element will know when there is gaps and allow users to retry fetching that part of the timeline.

Given the federation outage, it's very possible that the HS fell behind very far, then you paginated /messages and Synapse only backfilled 100 messages while you scrolled past that 2-day area leaving the gap.

Clearing the cache works because you re-paginate over /messages and Synapse has a second chance to get it right.

@MadLittleMods MadLittleMods added the O-Uncommon Most users are unlikely to come across this or unexpected workflow label Aug 25, 2022
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
A-Timeline O-Uncommon Most users are unlikely to come across this or unexpected workflow P1 S-Major Severely degrades major functionality or product features, with no satisfactory workaround T-Defect
Projects
None yet
Development

No branches or pull requests

3 participants