[SERVER-12400] --use-system-all adds -ltcmalloc to libmongclient.so Created: 17/Jan/14 Updated: 14/Aug/14 Resolved: 06/Aug/14 |
|
| Status: | Closed |
| Project: | Core Server |
| Component/s: | Build |
| Affects Version/s: | 2.4.6 |
| Fix Version/s: | None |
| Type: | Bug | Priority: | Major - P3 |
| Reporter: | Bruce Lucas (Inactive) | Assignee: | Unassigned |
| Resolution: | Won't Fix | Votes: | 1 |
| Labels: | None | ||
| Remaining Estimate: | Not Specified | ||
| Time Spent: | Not Specified | ||
| Original Estimate: | Not Specified | ||
| Issue Links: |
|
||||||||
| Participants: | |||||||||
| Description |
|
This causes libmongoclient.so to link tcmalloc, which should only be done from the main program. This has been observed to cause breakage after prelinking an app with the libmongoclient.so installed from the EPEL libmongodb-2.4.6-1.el6.x86_64 package. |
| Comments |
| Comment by Andrew Morrow (Inactive) [ 14/Aug/14 ] |
|
wholevinski FYI - |
| Comment by Andrew Morrow (Inactive) [ 13/Aug/14 ] |
|
wholevinski Actually, I spoke too soon. I just checked the 26compat branch and this issue is still present there. I've filed |
| Comment by Wesley Holevinski [ 13/Aug/14 ] |
|
Thanks, Andrew. This clears things up. |
| Comment by Andrew Morrow (Inactive) [ 13/Aug/14 ] |
|
Hi wholevinski - Sure. First, to be very clear: the C++ driver is not being dropped. Instead, the C++ driver has been forked from the server repository and is being developed independently in this repository: https://github.com/mongodb/mongo-cxx-driver Here is the blog post that provides an overview of why we have made this change and what the roadmap for the new driver looks like: http://blog.mongodb.org/post/78478565493/upcoming-changes-to-the-mongodb-cpp-driver The issue with the driver being inappropriately linked against tcmalloc has been resolved in both the 26compat release stream (currently issuing stable releases in lockstep with server 2.6.x releases), and in the legacy release stream (currently at release-candidate status, with a feature complete 1.0 anticipated later this summer). As far as status with EPEL, I don't have any particular insight into what they are doing. If you have information about packaging issues on EPEL, please let us know and we will try to coordinate with EPEL maintainers to resolve them. |
| Comment by Wesley Holevinski [ 13/Aug/14 ] |
|
@Andrew Morrow: Can you elaborate on this a little bit? Does that mean the C++ driver is being dropped? Or has it just been removed from EPEL and the problem with tcmalloc being included when compiling the SO still exists? |
| Comment by Andrew Morrow (Inactive) [ 06/Aug/14 ] |
|
No longer relevant with C++ driver removed from server repo. |