Open
Bug 1296572
Opened 8 years ago
Updated 2 years ago
Firefox UI slows down considerably after use due to DOM storage
Categories
(Core :: DOM: Core & HTML, defect, P3)
Tracking
()
UNCONFIRMED
People
(Reporter: joriswu, Unassigned)
References
Details
(Keywords: perf)
Attachments
(1 obsolete file)
User Agent: Mozilla/5.0 (X11; Linux x86_64; rv:47.0) Gecko/20100101 Firefox/47.0 Build ID: 20160623154057 Steps to reproduce: Install Firefox 47.0 On Linux Mint with blank profile, observe reasonable UI responsiveness. Use it for a month and observe sites like Google Maps and most others to slow down considerably. Scrolling becomes cumbersome and new pages take quite long to load. Actual results: Dependent on the site, opening new pages and scrolling becomes very slow. Mostly Javascript-rich and highly interactive sites suffer. webappsstore.sqlite is 16MB and webappsstore.sqlite-wal is 1.5MB Expected results: Performance should not degrade significantly over time.
See also bug 857888 which points to a possible underlying cause.
OS: Unspecified → Linux
Hardware: Unspecified → x86_64
After restarting with a single tab open on a trivial HTML-only webpage, VM usage is 1.5 GB (no extensions enabled) Responsiveness is still low. Removing the webappsstore.sqlite, and restarting again, the VM usage is 1.1 GB and UI responsiveness is improved. Cleaning the cache does not make a difference.
Updated•8 years ago
|
Comment 4•8 years ago
|
||
I don't know exactly where is the bottleneck in current local storage implementation, but I guess we should fix that as part of the next-gen local storage implementation.
Depends on: 1286798
Flags: needinfo?(jvarga)
Comment 5•8 years ago
|
||
Set this as P3 for now and we will review it after bug 1286798 is done.
Priority: -- → P3
I would just like to add that this is occurring for me too 51.0a2 (2016-10-04) (64-bit) Linux Mint I have noticed that once the webappsstore.sqlite gets large? (20mb or larger) the Firefox UI gets very choppy and unresponsive at times. I especially notice it for SPA's and other heavy JS sites. AWS Console is one site I have this problem on. Once I delete that file... Firefox works super well, and the UI is very responsive. https://www.reddit.com/r/firefox/comments/40gbos/psa_large_webappsstoresqlite_file_in_your_profile/
I've been noticing this as a user on Windows as well. Keep a long-term session open with lots of tabs, and then switch between remote desktop and native usage (I think the graphics switching from native to software rendering is partly the culprit here) at least a couple of times will cause Firefox to gradually slow down. Please let me know what else I can do to help narrow this down.
Comment 8•8 years ago
|
||
Could anyone who can reproduce this provide a cleopatra profile? Hopefully using Nightly, but if that is not possible, release version is fine too (it just gives less information). https://developer.mozilla.org/en-US/docs/Mozilla/Performance/Profiling_with_the_Built-in_Profiler
Comment 10•8 years ago
|
||
I think I caught a freeze in this profile (it certainly looks like it from the activity), but I may have lost some data from profiling too long.
Comment 11•7 years ago
|
||
Comment on attachment 8803654 [details] Firefox global freeze Sorry it took so long to take a look here! In the future, it would make it a bit easier to use if you used the Share button in the profiler UI to get a link. I've done that here: https://perfht.ml/2pLnBdp. A quick look through the profile seems to indicate that your computer is being slowed down outside Firefox. For example, it looks like the samples (which should probably be taken every 1 ms) are only being taken every 14 ms. That combined with some of the indications of slow painting, etc. indicate maybe your computer was under heavy load. Could this be the case? If anyone can reliably reproduce with a set of steps, please post them here. Again, apologies for the delay here.
Attachment #8803654 -
Attachment is obsolete: true
Assignee | ||
Updated•5 years ago
|
Component: DOM → DOM: Core & HTML
Updated•2 years ago
|
Severity: normal → S3
You need to log in
before you can comment on or make changes to this bug.
Description
•