Hi all, during debugging of an MDS problem we observed something that looks odd. The output of perf top seems to show symbols from v15 (octopus) on a pacific (v16) installation: 23.56% ceph-mds [.] std::_Rb_tree<inodeno_t, inodeno_t, std::_Identity<inodeno_t> 7.02% libceph-common.so.2 [.] ceph::buffer::v15_2_0::ptr::copy_out 4.99% ceph-mds [.] std::_Hashtable<inodeno_t, std::pair<inodeno_t const, CInode* 4.69% libceph-common.so.2 [.] ceph::buffer::v15_2_0::list::iterator_impl<true>::copy 2.53% ceph-mds [.] std::_Rb_tree<SnapRealm*, SnapRealm*, std::_Identity<SnapReal 2.46% libtcmalloc.so.4.5.3 [.] tcmalloc::CentralFreeList::Populate 2.31% [kernel] [k] _raw_spin_lock 2.11% ceph-mds [.] std::_Rb_tree<dentry_key_t, std::pair<dentry_key_t const, CDe 1.90% libstdc++.so.6.0.25 [.] std::_Rb_tree_insert_and_rebalance 1.86% [kernel] [k] copy_user_enhanced_fast_string 1.83% libtcmalloc.so.4.5.3 [.] operator new[] 1.83% libceph-common.so.2 [.] ceph::buffer::v15_2_0::list::iterator_impl<true>::operator+= 1.68% ceph-mds [.] MDCache::populate_mydir 1.56% ceph-mds [.] std::_Rb_tree<frag_t, std::pair<frag_t const, CDir*>, std::_S 1.38% [kernel] [k] clear_page_erms 1.06% [kernel] [k] native_irq_return_iret The pacific packages are installed from download.ceph.com and all package candidates claim to be v16: # yum provides "/*/libceph-common.so.2" Last metadata expiration check: 0:45:57 ago on Wed 15 Jan 2025 12:14:09 PM CET. librados2-2:16.2.15-0.el8.x86_64 : RADOS distributed object store client library Repo : @System Matched from: Filename : /usr/lib64/ceph/libceph-common.so.2 librados2-2:16.2.15-0.el8.x86_64 : RADOS distributed object store client library Repo : ceph Matched from: Filename : /usr/lib64/ceph/libceph-common.so.2 Why do we see v15 symbols here or am I interpreting the symbol name ceph::buffer::v15_2_0::list::iterator_impl<true>::copy incorrectly? Thanks and best regards, ================= Frank Schilder AIT Risø Campus Bygning 109, rum S14 _______________________________________________ ceph-users mailing list -- ceph-users@xxxxxxx To unsubscribe send an email to ceph-users-leave@xxxxxxx
 |