Ceph : Issueshttps://tracker.ceph.com/https://tracker.ceph.com/favicon.ico2023-09-27T13:55:27ZCeph
Redmine rgw - Bug #63006 (Closed): Unwatch crash on error at RGW startuphttps://tracker.ceph.com/issues/630062023-09-27T13:55:27ZAdam Emersonaemerson@redhat.com
<p>During radosgw initialization, if there is an exception in init_watch that causes the watcher registration to fail, When finalize_watch is executed, a crash occurs due to unregister an unregistered watch.</p> Ceph - Bug #62804 (New): Jaegertracing compile failurehttps://tracker.ceph.com/issues/628042023-09-11T16:48:07ZAdam Emersonaemerson@redhat.com
<p>Compilation fails with:</p>
<pre><code class="text syntaxhl"><span class="CodeRay">In file included from /home/aemerson/work/ceph/main/src/jaegertracing/opentelemetry-cpp/exporters/jaeger/src/TUDPTransport.h:16,
from /home/aemerson/work/ceph/main/src/jaegertracing/opentelemetry-cpp/exporters/jaeger/src/TUDPTransport.cc:6:
/home/aemerson/work/ceph/main/src/jaegertracing/opentelemetry-cpp/exporters/jaeger/src/TUDPTransport.cc: In member function ‘virtual void opentelemetry::v1::exporter::jaeger::TUDPTransport::close()’:
/home/aemerson/work/ceph/main/src/jaegertracing/opentelemetry-cpp/exporters/jaeger/src/TUDPTransport.cc:71:7: error: ‘::close’ has not been declared; did you mean ‘pclose’?
71 | ::THRIFT_CLOSESOCKET(socket_);
| ^~~~~~~~~~~~~~~~~~
</span></code></pre>
<p>It can be fixed with:</p>
<pre><code class="cpp syntaxhl"><span class="CodeRay">diff --git a/exporters/jaeger/src/TUDPTransport.cc b/exporters/jaeger/src/TUDPTransport.cc
index e4111273.<span class="float">.0</span>ea86288 <span class="integer">100644</span>
--- a/exporters/jaeger/src/TUDPTransport.cc
+++ b/exporters/jaeger/src/TUDPTransport.cc
<span class="error">@</span><span class="error">@</span> -<span class="integer">3</span>,<span class="integer">6</span> +<span class="integer">3</span>,<span class="integer">8</span> <span class="error">@</span><span class="error">@</span>
<span class="preprocessor">#include</span> <span class="include"><sstream></span> <span class="comment">// std::stringstream</span>
+<span class="preprocessor">#include</span> <span class="include"><unistd.h></span>
+
<span class="preprocessor">#include</span> <span class="include">"TUDPTransport.h"</span>
<span class="preprocessor">#include</span> <span class="include">"opentelemetry/sdk_config.h"</span>
</span></code></pre>
<p>applied to jaegertracing/opentelemetry-cpp</p> Ceph - Bug #62097 (Resolved): Quincy and pacific fail to compile after Boost upgrade of mainhttps://tracker.ceph.com/issues/620972023-07-20T01:20:22ZAdam Emersonaemerson@redhat.com
<p>After upgrading Boost on main, the Jenkins `make check` job fails on both quincy and pacific. This is caused by two issues:</p>
<p>1. install-deps on Quincy and Pacific was never updated after the switch to Jammy and it wouldn't surprise me if it'd been building against the wrong packages for a while now.</p>
<p>2. Boost.Phoenix 1.81 introduces an ODR-violation. A workaround for it is included in main, but not in anything earlier. It's probably worth including anyway in case someone tries to compile on a newer boost.</p> sepia - Bug #57186 (Closed): Please fix Folio11, which won't boothttps://tracker.ceph.com/issues/571862022-08-18T20:50:14ZAdam Emersonaemerson@redhat.com
<p>I seem to have broken something when trying to upgrade to Centos-Stream-9 to get a newer build toolchain, and I don't think I have the IPMI access needed to fix an unbooting machine.</p>
<p>Could someone fix this? If it's easiest just reimaging with new Fedora, CentOS-9-Stream, or Jammy would be ideal.</p>
<p>(But if that's not an option I'll take whatever's doable.)</p>
<p>Thank you.</p> rgw - Bug #57063 (Resolved): Incorrect resumption from Full to Incremental Sync in RGWDataSyncSha...https://tracker.ceph.com/issues/570632022-08-08T19:43:33ZAdam Emersonaemerson@redhat.com
<p>Because the `switch` statement is outside a resume block, if Full sync yields after the switch to incremental, the remaining code in `full_sync()` is never executed.</p> sepia - Bug #56665 (Resolved): Cobbler access requesthttps://tracker.ceph.com/issues/566652022-07-21T18:49:18ZAdam Emersonaemerson@redhat.com
<p>Please create an account for the Web cobbler UI for me.</p>
<p>Alternatively, please reimage Folio11, the system is old/unmaintained and hooked into dead CentOS repositories.</p> devops - Bug #56411 (Closed): Workaround for ceph-mgr breaks Cython buildshttps://tracker.ceph.com/issues/564112022-06-28T20:37:53ZAdam Emersonaemerson@redhat.com
<pre><code class="text syntaxhl"><span class="CodeRay">FAILED: lib/cython_modules/lib.3/rados.cpython-310-x86_64-linux-gnu.so /home/aemerson/work/ceph/repo/build/lib/cython_modules/lib.3/rados.cpython-310-x86_64-linux-gnu.so
cd /home/aemerson/work/ceph/repo/src/pybind/rados && env CC="/usr/bin/cc" CFLAGS="-fno-var-tracking-assignments" CPPFLAGS="-iquote/home/aemerson/work/ceph/repo/src/include -w -D'void0=dead_function(void)' -D'__Pyx_check_single_interpreter(ARG)=ARG ## 0'" CXX="/usr/bin/c++" LDSHARED="/usr/bin/cc -shared" OPT="-DNDEBUG -g -fwrapv -O2 -w" LDFLAGS=-L/home/aemerson/work/ceph/repo/build/lib CYTHON_BUILD_DIR=/home/aemerson/work/ceph/repo/build/src/pybind/rados CEPH_LIBDIR=/home/aemerson/work/ceph/repo/build/lib /usr/bin/python3.10 /home/aemerson/work/ceph/repo/src/pybind/rados/setup.py build --build-base /home/aemerson/work/ceph/repo/build/lib/cython_modules --build-platlib /home/aemerson/work/ceph/repo/build/lib/cython_modules/lib.3
Compiling rados.pyx because it changed.
[1/1] Cythonizing rados.pyx
running build
running build_ext
building 'rados' extension
C compiler: /usr/bin/cc -Wno-unused-result -Wsign-compare -DNDEBUG -g -fwrapv -O2 -Wall -g -fstack-protector-strong -Wformat -Werror=format-security -g -fwrapv -O2 -fno-var-tracking-assignments -iquote/home/aemerson/work/ceph/repo/src/include -w -Dvoid0=dead_function(void) -D__Pyx_check_single_interpreter(ARG)=ARG ## 0 -fPIC
creating /home/aemerson/work/ceph/repo/build/lib/cython_modules/temp.linux-x86_64-3.10
creating /home/aemerson/work/ceph/repo/build/lib/cython_modules/temp.linux-x86_64-3.10/home
creating /home/aemerson/work/ceph/repo/build/lib/cython_modules/temp.linux-x86_64-3.10/home/aemerson
creating /home/aemerson/work/ceph/repo/build/lib/cython_modules/temp.linux-x86_64-3.10/home/aemerson/work
creating /home/aemerson/work/ceph/repo/build/lib/cython_modules/temp.linux-x86_64-3.10/home/aemerson/work/ceph
creating /home/aemerson/work/ceph/repo/build/lib/cython_modules/temp.linux-x86_64-3.10/home/aemerson/work/ceph/repo
creating /home/aemerson/work/ceph/repo/build/lib/cython_modules/temp.linux-x86_64-3.10/home/aemerson/work/ceph/repo/build
creating /home/aemerson/work/ceph/repo/build/lib/cython_modules/temp.linux-x86_64-3.10/home/aemerson/work/ceph/repo/build/src
creating /home/aemerson/work/ceph/repo/build/lib/cython_modules/temp.linux-x86_64-3.10/home/aemerson/work/ceph/repo/build/src/pybind
creating /home/aemerson/work/ceph/repo/build/lib/cython_modules/temp.linux-x86_64-3.10/home/aemerson/work/ceph/repo/build/src/pybind/rados
compile options: '-I/usr/include/python3.10 -I/usr/include/python3.10 -c'
extra options: '-Wno-unused-result -Wsign-compare -DNDEBUG -g -fwrapv -O2 -Wall -g -fstack-protector-strong -Wformat -Werror=format-security -g -fwrapv -O2 -fno-var-tracking-assignments -iquote/home/aemerson/work/ceph/repo/src/include -w -Dvoid0=dead_function(void) -D__Pyx_check_single_interpreter(ARG)=ARG ## 0 -Wno-unused-result -Wsign-compare -DNDEBUG -g -fwrapv -O2 -Wall -g -fstack-protector-strong -Wformat -Werror=format-security -g -fwrapv -O2'
cc: /home/aemerson/work/ceph/repo/build/src/pybind/rados/rados.c
cc: warning: ##: linker input file unused because linking not done
cc: error: ##: linker input file not found: No such file or directory
cc: warning: 0: linker input file unused because linking not done
cc: error: 0: linker input file not found: No such file or directory
error: Command "/usr/bin/cc -Wno-unused-result -Wsign-compare -DNDEBUG -g -fwrapv -O2 -w -g -fstack-protector-strong -Wformat -Werror=format-security -g -fwrapv -O2 -fno-var-tracking-assignments -iquote/home/aemerson/work/ceph/repo/src/include -w -Dvoid0=dead_function(void) -D__Pyx_check_single_interpreter(ARG)=ARG ## 0 -fPIC -I/usr/include/python3.10 -I/usr/include/python3.10 -c /home/aemerson/work/ceph/repo/build/src/pybind/rados/rados.c -o /home/aemerson/work/ceph/repo/build/lib/cython_modules/temp.linux-x86_64-3.10/home/aemerson/work/ceph/repo/build/src/pybind/rados/rados.o -Wno-unused-result -Wsign-compare -DNDEBUG -g -fwrapv -O2 -Wall -g -fstack-protector-strong -Wformat -Werror=format-security -g -fwrapv -O2 -fno-var-tracking-assignments -iquote/home/aemerson/work/ceph/repo/src/include -w -Dvoid0=dead_function(void) -D__Pyx_check_single_interpreter(ARG)=ARG ## 0 -Wno-unused-result -Wsign-compare -DNDEBUG -g -fwrapv -O2 -Wall -g -fstack-protector-strong -Wformat -Werror=format-security -g -fwrapv -O2" failed with exit status 1
[818/1486] Building CXX object src/tools/ceph-dencoder/CMakeFiles/denc-mod-common.dir/common_types.cc.o
ninja: build stopped: subcommand failed.
</span></code></pre>
<p>Something is breaking the spacing here. This is with Python 3.10.</p>
<p>I get the same error whether I use make or ninja.</p>
<p>gcc (Debian 11.3.0-3) 11.3.0</p>
<p>/bin/sh is bash.</p>
<p>Cython version 0.29.30</p>
<p>This is directly from the bit of magic in Distutils.cmake that wipes out __Pyx_check_single_interpreter() , since if I remove that everything works.</p>
<p>(But I don't want to break ceph-mgr, so I can't really commit that.)</p> rgw - Bug #54208 (Resolved): Data race in RGWDataChangesLog::ChangeStatushttps://tracker.ceph.com/issues/542082022-02-08T18:36:45ZAdam Emersonaemerson@redhat.com
<p>Write to ChangeStatus while holding the wrong lock.</p> rgw - Bug #53941 (Resolved): Trim thread crashes when cluster misconfigured to have empty endpointshttps://tracker.ceph.com/issues/539412022-01-19T21:48:06ZAdam Emersonaemerson@redhat.com
<p>A zone with empty endpoints should print an error and bail out of the sync thread rather than crashing.</p> rgw - Bug #53325 (Resolved): Test failures in ceph_test_cls_rgw_gchttps://tracker.ceph.com/issues/533252021-11-18T20:09:52ZAdam Emersonaemerson@redhat.com
<p>It looks like this test was left out of teuthology by mistake and something broke somewhere along the line:</p>
<p>On a vstart cluster of one mmonitor and one OSD:</p>
<pre>
./bin/ceph_test_cls_rgw_gc
Running main() from gmock_main.cc
[==========] Running 11 tests from 1 test suite.
[----------] Global test environment set-up.
[----------] 11 tests from cls_rgw_gc
[ RUN ] cls_rgw_gc.init
[ OK ] cls_rgw_gc.init (1097 ms)
[ RUN ] cls_rgw_gc.gc_queue_ops1
[ ] list info tag = chain-0
[ OK ] cls_rgw_gc.gc_queue_ops1 (85 ms)
[ RUN ] cls_rgw_gc.gc_queue_ops2
[ ] next_marker is: = 0/4263
[ OK ] cls_rgw_gc.gc_queue_ops2 (40 ms)
[ RUN ] cls_rgw_gc.gc_queue_ops3
/home/azure/work/ceph/repo/src/test/cls_rgw_gc/test_cls_rgw_gc.cc:264: Failure
Expected equality of these values:
2
list_info1.size()
Which is: 1
[ FAILED ] cls_rgw_gc.gc_queue_ops3 (58 ms)
[ RUN ] cls_rgw_gc.gc_queue_ops4
/home/azure/work/ceph/repo/src/test/cls_rgw_gc/test_cls_rgw_gc.cc:343: Failure
Expected equality of these values:
2
list_info1.size()
Which is: 1
[ FAILED ] cls_rgw_gc.gc_queue_ops4 (58 ms)
[ RUN ] cls_rgw_gc.gc_queue_ops5
[ OK ] cls_rgw_gc.gc_queue_ops5 (61 ms)
[ RUN ] cls_rgw_gc.gc_queue_ops6
[ OK ] cls_rgw_gc.gc_queue_ops6 (54 ms)
[ RUN ] cls_rgw_gc.gc_queue_ops7
[ OK ] cls_rgw_gc.gc_queue_ops7 (50 ms)
[ RUN ] cls_rgw_gc.gc_queue_ops8
[ OK ] cls_rgw_gc.gc_queue_ops8 (57 ms)
[ RUN ] cls_rgw_gc.gc_queue_ops9
[ OK ] cls_rgw_gc.gc_queue_ops9 (60 ms)
[ RUN ] cls_rgw_gc.finalize
[ OK ] cls_rgw_gc.finalize (514 ms)
[----------] 11 tests from cls_rgw_gc (2134 ms total)
[----------] Global test environment tear-down
[==========] 11 tests from 1 test suite ran. (2134 ms total)
[ PASSED ] 9 tests.
[ FAILED ] 2 tests, listed below:
[ FAILED ] cls_rgw_gc.gc_queue_ops3
[ FAILED ] cls_rgw_gc.gc_queue_ops4
2 FAILED TESTS
</pre>
<p>This also reproduces on Pacific and Octopus.</p> rgw - Bug #53132 (Resolved): Old enough buckets may end up with an empty layout loghttps://tracker.ceph.com/issues/531322021-11-02T17:06:32ZAdam Emersonaemerson@redhat.com
<p>Pre-rgw::BucketLayout buckets will have most entries of rgw::BucketLayout synthesized, but the logs will be left empty.</p> rgw - Bug #51712 (Resolved): radosgw-admin should print error on missing --end-marker argumenthttps://tracker.ceph.com/issues/517122021-07-16T14:27:02ZAdam Emersonaemerson@redhat.com
<p>Rather than crashing.</p> rgw - Bug #51674 (Resolved): Potential race condition in robust notifyhttps://tracker.ceph.com/issues/516742021-07-14T14:04:59ZAdam Emersonaemerson@redhat.com
<p>Notifications can fail. This we know.</p>
<p>When they do, rather than repeating the update, which can lead to A-B-A style races, we should blow the object out of peer cache.</p> rgw - Bug #51661 (Resolved): Manual datalog trim segfaaultshttps://tracker.ceph.com/issues/516612021-07-13T20:02:58ZAdam Emersonaemerson@redhat.com
<p>If the target generation is the last, falls off the end of the map and tries to derefrence the past-the-end iterator.</p>
<p>Slipped through the testing since multisite tests always use the AioCompletion version.</p> rgw - Bug #48940 (Duplicate): Listing a new, empty omap datalog errorshttps://tracker.ceph.com/issues/489402021-01-20T16:42:11ZAdam Emersonaemerson@redhat.com
<p>On −ENOENT, we should return 0 and no entries, not −ENOENT.</p>