Open Bug 1765085 Opened 4 years ago Updated 2 years ago

Crash in [@ AsyncShutdownTimeout | places.sqlite#1: waiting for clients | places.sqlite#1: Bookmarks.jsm: fetchBookmark (34),places.sqlite#1: Bookmarks.jsm: fetchBookmark (37),places.sqlite#1: Bookmarks.jsm: fetchBookmarksByURL (33),places.sqlite#1: Hi...

Categories

(Toolkit :: Places, defect, P3)

Product:
Toolkit
Component:
Places
Version:
Firefox 99
Platform:
x86_64
Windows 10
Type:
defect

Tracking

()

People

(Reporter: martin.1980, Unassigned)

References

Details

Attachments

(1 file)

Integrity Check output
4.50 KB, text/plain
Details

Maybe Fission related. (DOMFissionEnabled=1)

Crash report: https://crash-stats.mozilla.org/report/index/af837851-c891-430a-b889-ac6a40220417

MOZ_CRASH Reason: [Parent 8548, Main Thread] ###!!! ABORT: file resource://gre/modules/Sqlite.jsm:471

Top 10 frames of crashing thread:

0 xul.dll NS_DebugBreak xpcom/base/nsDebugImpl.cpp:433
1 xul.dll nsDebugImpl::Abort xpcom/base/nsDebugImpl.cpp:132
2 xul.dll XPTC__InvokebyIndex 
3 None @0x000001c2e63bf81f 
4 xul.dll _tailMerge_d3dcompiler_47.dll 
5 xul.dll _tailMerge_d3dcompiler_47.dll 
6 xul.dll static XPCWrappedNative::CallMethod js/xpconnect/src/XPCWrappedNative.cpp:1130
7 xul.dll XPC_WN_CallMethod js/xpconnect/src/XPCWrappedNativeJSOps.cpp:923
8 xul.dll Interpret js/src/vm/Interpreter.cpp:3309
9 xul.dll js::Call js/src/vm/Interpreter.cpp:589

This bug occurs randomly on shutdown of FF and requests to send a bug report. Since no bug with this signature existed, I filed this one now.

The bug has a crash signature, thus the bug will be considered confirmed.

Status: UNCONFIRMED → NEW
Ever confirmed: true

One more info, it MIGHT be related to the number of bookmarks. I recently came across the border from 999 to more than 1000 bookmarks and it felt with me that after the > 1000 the browser started to hang e.g. also on adding new bookmarks (bookmark icon did not work, shortcut "add bookmark" did not work)

I don't think there's a strict relation to such a small number of bookmarks at that level, 1k bookmarks is nothing, we have users with tens thousands. There should also not be a relation with fission, that's very unlikely.
There may be other reasons though, a very busy mechanical disk, a corruption, a particularly expensive AV software, and more.

Let's start with simple checks, could you please execute Integrity Check from the Places Database section of about:support and post the resulting log?

Flags: needinfo?(martin.1980)
Attached file Integrity Check output —
Here is the log output:

Here is the log output:

Task: checkIntegrity

  • The places.sqlite database is sane
  • The favicons.sqlite database is sane

Task: invalidateCaches

  • The caches have been invalidated

Task: checkCoherence

  • The database is coherent

Task: expire

  • Database cleaned up

Task: originFrecencyStats

  • Recalculated origin frecency stats

Task: vacuum

  • Initial database size is 107520KiB
  • The database has been vacuumed
  • Final database size is 102400KiB

Task: stats

  • Places.sqlite size is 102400KiB
  • Favicons.sqlite size is 48960KiB
  • pragma_user_version is 64
  • pragma_page_size is 4096
  • pragma_cache_size is -2048
  • pragma_journal_mode is wal
  • pragma_synchronous is 1
  • History can store a maximum of 62714 unique pages
  • Table moz_bookmarks has 1093 records
  • Table moz_keywords has 1 records
  • Table sqlite_sequence has 1 records
  • Table moz_annos has 920 records
  • Table moz_anno_attributes has 3 records
  • Table moz_items_annos has 0 records
  • Table moz_places has 84918 records
  • Table moz_historyvisits has 880974 records
  • Table moz_inputhistory has 140 records
  • Table sqlite_stat1 has 20 records
  • Table moz_hosts has 0 records
  • Table moz_bookmarks_deleted has 5 records
  • Table moz_meta has 5 records
  • Table moz_origins has 5661 records
  • Table moz_places_metadata has 0 records
  • Table moz_places_metadata_search_queries has 0 records
  • Table moz_places_metadata_snapshots has 0 records
  • Table moz_places_metadata_snapshots_extra has 0 records
  • Table moz_places_metadata_snapshots_groups has 0 records
  • Table moz_places_metadata_groups_to_snapshots has 0 records
  • Table moz_session_metadata has 0 records
  • Table moz_session_to_places has 0 records
  • Table moz_previews_tombstones has 0 records
  • Pages sequentiality: 6.8
  • sqlite_schema: 16KiB (0%), 62% eff.
  • moz_keywords: 4KiB (0%), 0.5% eff.
  • sqlite_autoindex_moz_keywords_1: 4KiB (0%), 0.4% eff.
  • sqlite_sequence: 4KiB (0%), 0.7% eff.
  • moz_anno_attributes: 4KiB (0%), 2.5% eff.
  • sqlite_autoindex_moz_anno_attributes_1: 4KiB (0%), 2.5% eff.
  • moz_items_annos: 4KiB (0%), 0.2% eff.
  • moz_inputhistory: 4KiB (0%), 91.7% eff.
  • sqlite_autoindex_moz_inputhistory_1: 4KiB (0%), 64.4% eff.
  • sqlite_stat1: 4KiB (0%), 27.6% eff.
  • moz_hosts: 4KiB (0%), 0.2% eff.
  • sqlite_autoindex_moz_hosts_1: 4KiB (0%), 0.2% eff.
  • moz_bookmarks_deleted: 4KiB (0%), 3.2% eff.
  • sqlite_autoindex_moz_bookmarks_deleted_1: 4KiB (0%), 2.5% eff.
  • moz_meta: 4KiB (0%), 11.2% eff.
  • moz_places_metadata: 4KiB (0%), 0.2% eff.
  • moz_places_metadata_search_queries: 4KiB (0%), 0.2% eff.
  • sqlite_autoindex_moz_places_metadata_search_queries_1: 4KiB (0%), 0.2% eff.
  • moz_places_metadata_snapshots: 4KiB (0%), 0.2% eff.
  • moz_places_metadata_snapshots_extra: 4KiB (0%), 0.2% eff.
  • moz_places_metadata_snapshots_groups: 4KiB (0%), 0.2% eff.
  • moz_places_metadata_groups_to_snapshots: 4KiB (0%), 0.2% eff.
  • moz_session_metadata: 4KiB (0%), 0.2% eff.
  • sqlite_autoindex_moz_session_metadata_1: 4KiB (0%), 0.2% eff.
  • moz_session_to_places: 4KiB (0%), 0.2% eff.
  • moz_previews_tombstones: 4KiB (0%), 0.2% eff.
  • moz_bookmarks_itemindex: 20KiB (0%), 63.5% eff.
  • moz_bookmarks_parentindex: 20KiB (0%), 63.1% eff.
  • moz_bookmarks_itemlastmodifiedindex: 28KiB (0%), 75.7% eff.
  • moz_annos_placeattributeindex: 20KiB (0%), 60.4% eff.
  • moz_items_annos_itemattributeindex: 4KiB (0%), 0.2% eff.
  • moz_bookmarks_guid_uniqueindex: 28KiB (0%), 76.4% eff.
  • moz_keywords_placepostdata_uniqueindex: 4KiB (0%), 0.5% eff.
  • moz_bookmarks_dateaddedindex: 24KiB (0%), 71.3% eff.
  • moz_places_metadata_placecreated_uniqueindex: 4KiB (0%), 0.2% eff.
  • moz_places_metadata_referrerindex: 4KiB (0%), 0.2% eff.
  • moz_places_metadata_snapshots_pinnedindex: 4KiB (0%), 0.2% eff.
  • moz_places_metadata_snapshots_extra_typeindex: 4KiB (0%), 0.2% eff.
  • moz_bookmarks: 116KiB (0.1%), 93.2% eff.
  • moz_annos: 88KiB (0.1%), 90.9% eff.
  • moz_origins: 212KiB (0.2%), 97.1% eff.
  • sqlite_autoindex_moz_origins_1: 196KiB (0.2%), 97.7% eff.
  • moz_places_visitcount: 764KiB (0.8%), 99.2% eff.
  • moz_places_frecencyindex: 836KiB (0.8%), 99.6% eff.
  • moz_places_originidindex: 920KiB (0.9%), 99.4% eff.
  • moz_places_url_hashindex: 1260KiB (1.2%), 99% eff.
  • moz_places_lastvisitdateindex: 1420KiB (1.4%), 99.2% eff.
  • moz_places_guid_uniqueindex: 1760KiB (1.7%), 99.2% eff.
  • moz_places_hostindex: 2084KiB (2.1%), 99.2% eff.
  • moz_historyvisits_fromindex: 10292KiB (10.2%), 99.7% eff.
  • moz_historyvisits_dateindex: 14704KiB (14.5%), 99.8% eff.
  • moz_historyvisits_placedateindex: 18188KiB (17.9%), 99.6% eff.
  • moz_historyvisits: 23348KiB (23%), 99.7% eff.
  • moz_places: 24920KiB (24.6%), 95.7% eff.

Task: _refreshUI

Crash Signature: [@ AsyncShutdownTimeout | places.sqlite#1: waiting for clients | places.sqlite#1: Bookmarks.jsm: fetchBookmark (34),places.sqlite#1: Bookmarks.jsm: fetchBookmark (37),places.sqlite#1: Bookmarks.jsm: fetchBookmarksByURL (33),places.sqlite#1: History.jsm: i…
Flags: needinfo?(martin.1980)

The severity field is not set for this bug.
:mak, could you have a look please?

For more information, please visit auto_nag documentation.

Flags: needinfo?(mak)

Do you have any third-party software, such as anti-virus, or cache cleaners that could be interfering with Firefox? If so, please could you try temporarily disabling them and see if that helps.

Flags: needinfo?(mak) → needinfo?(martin.1980)

I use MS Defender only and no cleaning SW. One more topic i observed: it happened again with Firefox 100 (or 100.0.1) but only after the PC resumed from a "suspend to disc". And more strange - no crash report this time.
Feel free to request more details, I can provide a dump of the profile or whatever is required

Flags: needinfo?(martin.1980)

I tried playing around with plugins, but could not really figure out a dedicated reason.
But a new Crash report: https://crash-stats.mozilla.org/report/index/cdcba513-8fc9-449d-bd7b-3cc200220523

I now tried a new Profile, just copied tabs, bookmarks and pwds. I will update here in a few days

How did the new profile go?

Do you think there is a correlation with suspect to disc happening?

Severity: -- → S3
Component: Bookmarks & History → Places
Flags: needinfo?(martin.1980)
Priority: -- → P3
Product: Firefox → Toolkit

The new profile helped, but not for long.
Suspend to disc is still a potential troublemaker.
Anyhow, today a new crash report: https://crash-stats.mozilla.org/report/index/673b9ca9-af58-4b71-a293-7bc440220610

Flags: needinfo?(martin.1980)

This looks very similar to a bug we had in the past where Suspending/Hybernating the system was leaving Sqlite transactions in a pending state.
We fixed that in Bug 1727261 for Firefox 94 by adding a 5 minutes timeout and just rolling back the transaction after that.
It looks like there's still an unhandled case around.
From the crash it looks like all the statements are pending and waiting for something to happen and unblock them. It's likely worth a second check to the Sqlite.jsm timeout handling.

What you could try to do is set toolkit.sqlitejsm.loglevel to Warn in about:config, 5 minutes after suspension ends check in the Browser Console if you see any error message about transactions. Unfortunately it's not possible to get the log after the crash, so this will be a bit annoying to do.

See Also: → 1727261

Well, there are sme interesting messages:
[Exception... "Component returned failure code: 0x80040111 (NS_ERROR_NOT_AVAILABLE) [nsIContentSniffer.getMIMETypeFromContent]" nsresult: "0x80040111 (NS_ERROR_NOT_AVAILABLE)" location: "JS frame :: resource:///modules/FaviconLoader.jsm :: onStopRequest :: line 312" data: no] FaviconLoader.jsm:312:24
onStopRequest resource:///modules/FaviconLoader.jsm:312
InterpretGeneratorResume self-hosted:1422
AsyncFunctionNext self-hosted:632

Transaction timeout, most likely caused by unresolved pending work. Sqlite.jsm:1036
timeoutPromise resource://gre/modules/Sqlite.jsm:1036
notify resource://gre/modules/Timer.jsm:62

Duplicate of this bug: 1746773
You need to log in before you can comment on or make changes to this bug.

Attachment

General

Creator:
Created:
Updated:
Size: