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

The number of accounts/ files in TdS and SLP snapshots is too high #8168

Closed
mvines opened this issue Feb 7, 2020 · 7 comments · Fixed by #8482
Closed

The number of accounts/ files in TdS and SLP snapshots is too high #8168

mvines opened this issue Feb 7, 2020 · 7 comments · Fixed by #8482
Assignees
Milestone

Comments

@mvines
Copy link
Member

mvines commented Feb 7, 2020

Snapshots are taking 1000 slots or worse to unpack and process on TdS/SLP, due to the large number of accounts/ files in them. At this time, TdS has over 4500 files in accounts/, SLP has almost 3000.

This slows down the ability of validators to quickly boot and restart.

@mvines mvines added this to the Tofino v0.23.3 milestone Feb 7, 2020
@ryoqun
Copy link
Member

ryoqun commented Feb 10, 2020

This is partly affected by this #8148, which causes some unneeded storage entries not to be deleted forever.

@mvines
Copy link
Member Author

mvines commented Feb 15, 2020

This is starting to be a real problem. I'm getting snapshots on TdS and SLP that are from more than 10 minutes ago, which causes new validators to catch up very slowly

@ryoqun
Copy link
Member

ryoqun commented Feb 19, 2020

Status update: I've created a fix locally which reduces the current TdS snapshot into 1/5. PRs are coming shortly after.

@mvines
Copy link
Member Author

mvines commented Feb 27, 2020

oops, this is not fixed!

@mvines mvines reopened this Feb 27, 2020
@mvines mvines modified the milestones: Tofino v0.23.8, Tofino v0.23.9 Feb 29, 2020
@ryoqun
Copy link
Member

ryoqun commented Mar 13, 2020

Finally, my almost 1-month worth labor is paying off:

TdS:

image

SLP:

image

@mvines mvines modified the milestones: v1.0.7, v1.0.8 Mar 16, 2020
@mvines
Copy link
Member Author

mvines commented Mar 16, 2020

@ryoqun - I think we can call this issue fixed now right?

@ryoqun
Copy link
Member

ryoqun commented Mar 16, 2020

@mvines yes from operational standpoint!

There is still minor security issue, which I noticed and could be mitigated by more aggressive cleaning implementation. But there are bunch of similar low-priority security issues when deserializing a snapshot, which is tracked by #7167. So, closing this now.

@ryoqun ryoqun closed this as completed Mar 16, 2020
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
3 participants
@ryoqun @mvines and others