[SERVER-51476] Upgrade Abseil to 20210324.1 Created: 09/Oct/20  Updated: 29/Oct/23  Resolved: 13/Jul/21

Status: Closed
Project: Core Server
Component/s: Build
Affects Version/s: None
Fix Version/s: 5.1.0-rc0

Type: Task Priority: Major - P3
Reporter: Mark Benvenuto Assignee: Billy Donahue
Resolution: Fixed Votes: 0
Labels: servicearch-wfbf-day, third-party-libraries, third-party-upgrade
Remaining Estimate: Not Specified
Time Spent: Not Specified
Original Estimate: Not Specified

Issue Links:
Backports
Depends
depends on SERVER-57267 Dereference of invalid iterator in co... Closed
depends on SERVER-57477 stdx::erase_if for stdx::unordered_map Closed
depends on SERVER-57480 Invalid iterator access in api_versio... Closed
depends on SERVER-58263 Fix absl map erase loop in TenantMigr... Closed
depends on SERVER-58332 pipeline MemoryUsageTracker: change S... Closed
depends on SERVER-58370 StringMap element use-after-free in R... Closed
is depended on by SERVER-58675 Upgrade Abseil to next LTS release Open
Problem/Incident
causes SERVER-59107 Fix abseil pretty printers Closed
Related
related to SERVER-74523 Upgrade abseil to 2023-08-02 LTS Closed
Backwards Compatibility: Fully Compatible
Backport Requested:
v5.0
Sprint: Service Arch 2021-06-28, Service Arch 2021-07-12, Service Arch 2021-07-26
Participants:
Story Points: 3

 Description   

Abseil has issued a new 20210324.1 LTS release. We should evaluate whether we wish to upgrade

Acceptance criteria:

file tickets for all issues blocking upgrade. Create a new ticket to do the upgrade if we decide that it is useful. 

Figure out the latest LTS version for Abseil before starting this ticket.



 Comments   
Comment by Githook User [ 10/Dec/21 ]

Author:

{'name': 'Billy Donahue', 'email': 'billy.donahue@mongodb.com', 'username': 'BillyDonahue'}

Message: SERVER-51476 upgrade to Abseil 20210324.1

(cherry picked from commit defbe4582778e3da3abdc23c73ef7639543ab380)
Branch: v5.0
https://github.com/mongodb/mongo/commit/3845b16f0476ef8ea577b8eb09e34f23c76391f0

Comment by Vivian Ge (Inactive) [ 06/Oct/21 ]

Updating the fixversion since branching activities occurred yesterday. This ticket will be in rc0 when it’s been triggered. For more active release information, please keep an eye on #server-release. Thank you!

Comment by Billy Donahue [ 13/Jul/21 ]

We might want to upgrade to the .2 LTS patchlevel soon. I didn't want to retest with it as I feel the gains here in the .1 patchlevel are worth locking in asap.

Comment by Githook User [ 13/Jul/21 ]

Author:

{'name': 'Billy Donahue', 'email': 'billy.donahue@mongodb.com', 'username': 'BillyDonahue'}

Message: SERVER-51476 upgrade to Abseil 20210324.1
Branch: master
https://github.com/mongodb/mongo/commit/defbe4582778e3da3abdc23c73ef7639543ab380

Comment by Billy Donahue [ 13/Jul/21 ]

I'm gonna push. I checked with #BB and a reasonably green EVG patch seems to consist only of known issues. The acceptance criteria says I should make another ticket for the upgrade but I'm not doing that. The number SERVER-51476 is encoded into TODOs and stuff. Making a new ticket would dilute information.

Comment by Billy Donahue [ 08/Jul/21 ]

Another run, this time with all fixes up to and including SERVER-58370.
https://evergreen.mongodb.com/version/60e778d4c9ec440e556d96cf

Comment by Billy Donahue [ 08/Jul/21 ]

A fix is pending for SERVER-58332. StringMap is a flat_hash_map, not a node_hash_map. This difference matters more now I guess.
Another run of this upgrade with that fix in place:
https://spruce.mongodb.com/version/60e65d07d6d80a48fd8baf07

Comment by Andrew Morrow (Inactive) [ 04/Jun/21 ]

Thanks, I'll get my branch rebased and give it another run.

Comment by George Wangensteen [ 27/May/21 ]

acm I'm planning on taking a deeper look at this next sprint if you want to hold off until then.

Comment by George Wangensteen [ 27/May/21 ]

You could give it a try, but I would consider holding off for now: we haven't yet done a thorough investigation of your patch build/taken a look at all the linked failures. I was looking at the particular failure you linked + its core dump as I thought it might be related to/offer a clue with an existing BF I had, which is how I found the bug I linked. It's hard to say how many of the failures in your build this is responsible for without looking into it further.

Given that we found a bug from that one failure, though, I think we (SA) should prioritize this work when possible (by "this work" I mean: looking through your patch build, and particularly the sanitizer failures, more throughly) both to see if it reveals any additional bugs on our end and to get in the abseil upgrade. This ticket is sitting near the top of our ordered backlog of non-project work, so hopefully we'll get to it soon. I'll bring it up in our next triage meeting as well. I'm curious if you wanted to get the upgrade in pre-5.0?

Comment by Andrew Morrow (Inactive) [ 27/May/21 ]

Should I rebase my abseil branch and try again after that ticket is resolved? Maybe we can merge it.

Comment by George Wangensteen [ 27/May/21 ]

Thanks for sending this our way, we already found at least one bug from investigating this (see the linked ticket).

Generated at Thu Feb 08 05:25:35 UTC 2024 using Jira 9.7.1#970001-sha1:2222b88b221c4928ef0de3161136cc90c8356a66.