Server-enforced ACLs by jdarcy · Pull Request #1 · gluster/glusterfs

jdarcy · 2011-04-19T13:27:19Z

This is the feature requested by one of my CloudFS/GlusterFS users. Commit comment follows...

By default, the behavior is the same as currently - ACLs are interpreted and
enforced on the client. To use server-side ACLs:

(1) Make sure that your server-side filesystems support the "acl" option and
are mounted with it.

(2) Add "option use-set-fsid true" to all of your server-side storage/posix
translators.

(3) Mount with "-o default-permissions=false" on the client side.

Note that the GlusterFS native protocol has no way to transmit supplementary
group IDs, so accesses relying on those might fail even if the ACL is set
correctly. This might be the subject of a future patch.

By default, the behavior is the same as currently - ACLs are interpreted and enforced on the client. To use server-side ACLs: (1) Make sure that your server-side filesystems support the "acl" option and are mounted with it. (2) Add "option use-set-fsid true" to all of your server-side storage/posix translators. (3) Mount with "-o default-permissions=false" on the client side. Note that the GlusterFS native protocol has no way to transmit supplementary group IDs, so accesses relying on those might fail even if the ACL is set correctly. This might be the subject of a future patch.

abperiasamy · 2011-04-19T17:56:21Z

This sounds great. I am discussing with Vijay and Avati to help you with necessary changes towards complete ACL support (like protocol change..).

avati · 2011-04-25T15:52:36Z

We are looking at supporting Posix ACLs with a native translator -

http://patches.gluster.com/patch/7040/
http://patches.gluster.com/patch/7041/
http://patches.gluster.com/patch/7042/
http://patches.gluster.com/patch/7043/
http://patches.gluster.com/patch/7044/
http://patches.gluster.com/patch/7047/

jdarcy · 2011-04-25T15:58:30Z

Which version do you expect will have that?

abperiasamy · 2011-04-25T16:07:52Z

Not decided on version yet.

avati · 2011-04-25T17:08:24Z

oops! merged by mistake :|

abperiasamy · 2011-04-25T17:09:54Z

I know that merge button was tempting :P

On Mon, Apr 25, 2011 at 10:08 AM, avati
reply@reply.github.com
wrote:

oops! merged by mistake :|

Reply to this email directly or view it on GitHub:
#1 (comment)

Anand Babu Periasamy
Blog [http://www.unlocksmith.org]

Imagination is more important than knowledge --Albert Einstein

jdarcy · 2011-04-26T13:02:32Z

Gee, I guess I should have tested that code. ;)

avati · 2011-04-26T13:08:21Z

I reverted it right after the accidental merge. The other ACL patches I
mentioned will be coming into mainline in a few days. Therefore it will
enter the next (3.3) release train (by default). Whether we backport it to
3.2 or 3.1, we're not sure yet.

glusterd: Add hostname to list instead of replaceing upon update

This patch improves the peer identification mechanism in glusterd and lays down the framework for further improvements, including better multi network support in glusterd. This patch mainly does two things, 1. Extend the peerinfo object to store a list of addresses instead of a single hostname as it does now. This also includes changes to make the peer update behaviour of 'peer probe' to add to the list. 2. Improve glusterd_friend_find_by_hostname() to perform better matching of hostnames. glusterd_friend_find_by_hostname() now does and initial quick string compare against all the peer addresses known to glusterd, after which it tries a more thorough search using address resolution and matching the struc sockaddr's. The above two changes together improve the peer identification situation in glusterd a lot. More information regarding the problem this patch attempts to resolve and the approach chosen can be found at http://www.gluster.org/community/documentation/index.php/Features/Better_peer_identification This commit is a squashed commit of the following changes, the development branch of which can be viewed at, https://github.com/kshlm/glusterfs/tree/better-peer-identification or, https://forge.gluster.org/~kshlm/glusterfs-core/kshlms-glusterfs/commits/better-peer-identification commit 198f86e Author: Kaushal M <kaushal@redhat.com> Date: Tue Jul 15 14:34:06 2014 +0530 Update gluster.8 commit 35d597f Author: Kaushal M <kaushal@redhat.com> Date: Tue Jul 15 09:01:01 2014 +0530 Address review comments https://review.gluster.org/#/c/8238/3 commit 47b5331 Merge: c71b12c 78128af Author: Kaushal M <kaushal@redhat.com> Date: Tue Jul 15 08:41:39 2014 +0530 Merge branch 'master' into better-peer-identification commit c71b12c Merge: 57bc9de 0f5719a Author: Kaushal M <kaushal@redhat.com> Date: Thu Jul 10 19:50:19 2014 +0530 Merge branch 'master' into better-peer-identification commit 57bc9de Author: Kaushal M <kaushal@redhat.com> Date: Thu Jul 10 19:49:08 2014 +0530 More fixes to review comments commit 5482cc3 Author: Kaushal M <kaushal@redhat.com> Date: Thu Jul 10 18:36:40 2014 +0530 Code refactoring in peer-utils based on review comments https://review.gluster.org/#/c/8238/2/xlators/mgmt/glusterd/src/glusterd-peer-utils.c commit 89b22c3 Author: Kaushal M <kaushal@redhat.com> Date: Thu Jul 10 12:30:00 2014 +0530 Hostnames in peer status commit 63ebf94 Merge: c8c8fdd f5f9721 Author: Kaushal M <kaushal@redhat.com> Date: Thu Jul 10 12:06:33 2014 +0530 Merge remote-tracking branch 'origin/master' into better-peer-identification commit c8c8fdd Author: Kaushal M <kaushal@redhat.com> Date: Wed Jul 9 18:35:27 2014 +0530 Hostnames in xml output commit 732a92a Author: Kaushal M <kaushal@redhat.com> Date: Wed Jul 9 15:12:10 2014 +0530 Add hostnames to cli rsp dict during list-friends commit fcf43e3 Merge: c0e2624 72d96e2 Author: Kaushal M <kaushal@redhat.com> Date: Wed Jul 9 12:53:03 2014 +0530 Merge branch 'master' into better-peer-identification commit c0e2624 Author: Kaushal M <kaushal@redhat.com> Date: Mon Jul 7 16:11:19 2014 +0530 Use list_for_each_entry_safe when cleaning peer hostnames commit 6132e60 Author: Kaushal M <kaushal@redhat.com> Date: Mon Jul 7 15:52:19 2014 +0530 Fix crash in gd_add_friend_to_dict commit 88ffa9a Author: Kaushal M <kaushal@redhat.com> Date: Mon Jul 7 13:19:44 2014 +0530 gd_peerinfo_destroy -> glusterd_peerinfo_destroy commit 4b36930 Author: Kaushal M <kaushal@redhat.com> Date: Mon Jul 7 12:50:12 2014 +0530 More refactoring commit ee559b0 Author: Kaushal M <kaushal@redhat.com> Date: Mon Jul 7 12:14:40 2014 +0530 Major refactoring of code based on review comments at https://review.gluster.org/#/c/8238/1/xlators/mgmt/glusterd/src/glusterd-peer-utils.h commit e96dbc7 Merge: 2613d1d 83c09b7 Author: Kaushal M <kaushal@redhat.com> Date: Mon Jul 7 09:47:05 2014 +0530 Merge remote-tracking branch 'origin/master' into better-peer-identification commit 2613d1d Merge: b242cf6 9a50211 Author: Kaushal M <kaushal@redhat.com> Date: Fri Jul 4 15:28:57 2014 +0530 Merge remote-tracking branch 'origin/master' into better-peer-identification commit b242cf6 Author: Kaushal M <kaushal@redhat.com> Date: Fri Jul 4 15:08:18 2014 +0530 Fix a silly mistake, if (ctx->req) => if (ctx->req == NULL) commit c835ed2 Author: Kaushal M <kaushal@redhat.com> Date: Fri Jul 4 14:58:23 2014 +0530 Fix reverse probe. commit 9ede17f Author: Kaushal M <kaushal@redhat.com> Date: Fri Jul 4 13:31:32 2014 +0530 Fix friend import for existing peers commit 891bf74 Author: Kaushal M <kaushal@redhat.com> Date: Fri Jul 4 13:08:36 2014 +0530 Set first hostname in peerinfo->hostnames to peerinfo->hostname commit 9421d6a Author: Kaushal M <kaushal@redhat.com> Date: Fri Jul 4 12:21:40 2014 +0530 Fix gf_asprintf return val check in glusterd_store_peer_write commit defac97 Author: Kaushal M <kaushal@redhat.com> Date: Fri Jul 4 11:16:13 2014 +0530 Fix store_retrieve_peers to correctly cleanup. commit 00a799f Author: Kaushal M <kaushal@redhat.com> Date: Fri Jul 4 10:52:11 2014 +0530 Update address list in glusterd_probe_cbk only when needed. commit 7a628e8 Merge: d191985 dc46d5e Author: Kaushal M <kaushal@redhat.com> Date: Fri Jul 4 09:24:12 2014 +0530 Merge remote-tracking branch 'origin/master' into better-peer-identification commit d191985 Author: Kaushal M <kaushal@redhat.com> Date: Tue Jul 1 18:59:49 2014 +0530 gf_compare_addrinfo -> gf_compare_sockaddr commit 31d8ef7 Merge: 93bbede 86ee233 Author: Kaushal M <kaushal@redhat.com> Date: Tue Jul 1 18:16:13 2014 +0530 Merge remote-tracking branch 'origin/master' into better-peer-identification commit 93bbede Author: Kaushal M <kaushal@redhat.com> Date: Tue Jul 1 18:15:16 2014 +0530 Improve glusterd_friend_find_by_hostname glusterd_friend_find_by_hostname will now do an initial quick search for the peerinfo performing string comparisions on the given host string. It follows it with a more thorough match, by resolving the addresses and comparing addrinfos instead of strings. commit 2542cdb Author: Kaushal M <kaushal@redhat.com> Date: Tue Jul 1 17:21:10 2014 +0530 New utility gf_compare_addrinfo commit 338676e Author: Kaushal M <kaushal@redhat.com> Date: Tue Jul 1 14:55:56 2014 +0530 Use gd_peer_has_address instead of strcmp commit 28d45be Merge: 728266e 991dd5e Author: Kaushal M <kaushal@redhat.com> Date: Tue Jul 1 14:54:40 2014 +0530 Merge branch 'master' into better-peer-identification commit 728266e Merge: 7d9b87b 2417de9 Author: Kaushal M <kaushal@redhat.com> Date: Tue Jul 1 09:55:13 2014 +0530 Merge remote-tracking branch 'origin/master' into better-peer-identification commit 7d9b87b Merge: b890625 e02275c Author: Kaushal M <kshlmster@gmail.com> Date: Tue Jul 1 08:41:40 2014 +0530 Merge pull request #4 from vpshastry/better-peer-identification Better peer identification commit e02275c Merge: 75ee90d b890625 Author: Varun Shastry <vshastry@redhat.com> Date: Mon Jun 30 16:44:29 2014 +0530 Merge branch 'better-peer-identification' of https://github.com/kshlm/glusterfs into better-peer-identification-kaushal-github commit 75ee90d Author: Varun Shastry <vshastry@redhat.com> Date: Mon Jun 30 15:36:10 2014 +0530 glusterd: add to the list if the probed uuid pre-exists Signed-off-by: Varun Shastry <vshastry@redhat.com> commit b890625 Merge: 04c5d60 187a7a9 Author: Kaushal M <kaushal@redhat.com> Date: Mon Jun 30 11:44:13 2014 +0530 Merge remote-tracking branch 'origin/master' into better-peer-identification commit 04c5d60 Merge: 3a5bfa1 e01edb6 Author: Kaushal M <kshlmster@gmail.com> Date: Sat Jun 28 19:23:33 2014 +0530 Merge pull request #3 from vpshastry/better-peer-identification glusterd: search through the list of hostnames in the peerinfo commit 0c64f33 Merge: e01edb6 3a5bfa1 Author: Varun Shastry <vshastry@redhat.com> Date: Sat Jun 28 10:43:29 2014 +0530 Merge branch 'better-peer-identification' of https://github.com/kshlm/glusterfs into better-peer-identification-kaushal-github commit e01edb6 Author: Varun Shastry <vshastry@redhat.com> Date: Fri Jun 27 12:29:36 2014 +0530 glusterd: search through the list of hostnames in the peerinfo Signed-off-by: Varun Shastry <vshastry@redhat.com> commit 3a5bfa1 Merge: cda6d31 371ea35 Author: Kaushal M <kshlmster@gmail.com> Date: Fri Jun 27 11:31:17 2014 +0530 Merge pull request #1 from vpshastry/better-peer-identification glusterd: Add hostname to list instead of replaceing upon update commit 371ea35 Author: Varun Shastry <vshastry@redhat.com> Date: Fri Jun 27 11:24:54 2014 +0530 glusterd: Add hostname to list instead of replaceing upon update Signed-off-by: Varun Shastry <vshastry@redhat.com> commit cda6d31 Author: Kaushal M <kaushal@redhat.com> Date: Thu Jun 26 19:52:52 2014 +0530 Import address lists commit 6649b54 Author: Kaushal M <kaushal@redhat.com> Date: Thu Jun 26 19:15:37 2014 +0530 Implement export address list commit 5599003 Author: Kaushal M <kaushal@redhat.com> Date: Thu Jun 26 18:11:59 2014 +0530 Use first address in list to when setting up the peer RPC. commit a35fde8 Author: Kaushal M <kaushal@redhat.com> Date: Thu Jun 26 18:03:04 2014 +0530 Properly free addresses on glusterd_peer_destroy commit 1988081 Author: Kaushal M <kaushal@redhat.com> Date: Thu Jun 26 17:52:35 2014 +0530 Restore peerinfo with address list implemented. commit 66f524d Author: Kaushal M <kaushal@redhat.com> Date: Mon Jun 23 13:02:23 2014 +0530 Move out all peer related utilities from glusterd-utils to glusterd-peer-utils commit 14a2a32 Author: Kaushal M <kaushal@redhat.com> Date: Tue May 27 12:16:41 2014 +0530 Compilation fix commit c59cd35 Author: Kaushal M <kaushal@redhat.com> Date: Mon May 5 12:51:11 2014 +0530 Add store support for hostname list commit b70325f Author: Kaushal M <kaushal@redhat.com> Date: Fri May 2 15:58:07 2014 +0530 Add a hostnames list to glusterd_peerinfo_t glusterd_peerinfo_new will now init this list and add the given hostname as the lists first member. Signed-off-by: Kaushal M <kaushal@redhat.com> Signed-off-by: Varun Shastry <vshastry@redhat.com> Change-Id: Ief3c5d6d6f16571ee2fab0a45e638b9d6506a06e BUG: 1119547 Reviewed-on: http://review.gluster.org/8238 Reviewed-by: Atin Mukherjee <amukherj@redhat.com> Tested-by: Gluster Build System <jenkins@build.gluster.com> Reviewed-by: Krishnan Parthasarathi <kparthas@redhat.com> Tested-by: Krishnan Parthasarathi <kparthas@redhat.com>

Resync

Coverity CID 1288822 (#1 of 2) strncpy executed with a limit equal to the target array size potentially leaves the target string not null terminated. In this case the strncpy is not needed due to the snprintf with the same target buffer which follows immediately. This patch also removes the now unneeded scratch_dir argument to gf_changelog_init_history(), which is semantically correct, since scratch_dir has previously been filled into jnl->jnl_working_dir by the caller, and this is now used to fill hist_scratch_dir. Change-Id: Ib1ed3a1058e80e34191758921b49c29030d6c9db BUG: 789278 Signed-off-by: Michael Adam <obnox@samba.org> Reviewed-on: http://review.gluster.org/10058 Reviewed-by: Kotresh HR <khiremat@redhat.com> Tested-by: Gluster Build System <jenkins@build.gluster.com> Reviewed-by: Vijay Bellur <vbellur@redhat.com>

up on failure case __inode_ctx_put need to free the allocated memory Indirect leak of 104 byte(s) in 1 object(s) allocated from: #0 0x496669 in __interceptor_calloc (/usr/local/sbin/glusterfsd+0x496669) #1 0x7f8a288522f9 in __gf_calloc libglusterfs/src/mem-pool.c:117 #2 0x7f8a17235962 in __posix_acl_ctx_get xlators/system/posix-acl/src/posix-acl.c:308 Change-Id: I0ce6da3967c55931a70f77d8551ccf52e4cdfda3 BUG: 1338733 Signed-off-by: Prasanna Kumar Kalever <prasanna.kalever@redhat.com> Reviewed-on: http://review.gluster.org/14505 Tested-by: Prasanna Kumar Kalever <pkalever@redhat.com> Smoke: Gluster Build System <jenkins@build.gluster.com> NetBSD-regression: NetBSD Build System <jenkins@build.gluster.org> CentOS-regression: Gluster Build System <jenkins@build.gluster.com> Reviewed-by: Jeff Darcy <jdarcy@redhat.com>

…ories Problem #1: Hitting CTRL+C leaves stale processes on remote nodes if glusterfind pre has been initiated. Solution #1: Adding "-t -t" to ssh command-line forces pseudo-terminal to be assigned to remote process. When local process receives Keyboard Interrupt, SIGHUP is immediately conveyed to the remote terminal causing remote changelog.py process to terminate immediately. Problem #2: Concurrent glusterfind pre runs are not possible on the same glusterfind session in case of a runaway process. Solution #2: glusterfind pre runs now add random directory name to the working directory to store and manage temporary database and changelog processing. If KeyboardInterrupt is received, the function call run_cmd_nodes("cleanup", args, tmpfilename=gtmpfilename) cleans up the remote run specific directory. Patch: 7571380 cli/xml: Fix wrong XML format in volume get command broke "gluster volume get <vol> changelog.rollover-time --xml" Now fixed function utils.py::get_changelog_rollover_time() Fixed spurious trailing space getting written if second path is empty in main.py::write_output() Fixed repetitive changelog processing in changelog.py::get_changes() Change-Id: Ia8d96e2cd47bf2a64416bece312e67631a1dbf29 BUG: 1382236 Signed-off-by: Milind Changire <mchangir@redhat.com> Reviewed-on: http://review.gluster.org/15609 Smoke: Gluster Build System <jenkins@build.gluster.org> CentOS-regression: Gluster Build System <jenkins@build.gluster.org> NetBSD-regression: NetBSD Build System <jenkins@build.gluster.org> Reviewed-by: Aravinda VK <avishwan@redhat.com>

…ories Problem #1: Hitting CTRL+C leaves stale processes on remote nodes if glusterfind pre has been initiated. Solution #1: Adding "-t -t" to ssh command-line forces pseudo-terminal to be assigned to remote process. When local process receives Keyboard Interrupt, SIGHUP is immediately conveyed to the remote terminal causing remote changelog.py process to terminate immediately. Problem #2: Concurrent glusterfind pre runs are not possible on the same glusterfind session in case of a runaway process. Solution #2: glusterfind pre runs now add random directory name to the working directory to store and manage temporary database and changelog processing. If KeyboardInterrupt is received, the function call run_cmd_nodes("cleanup", args, tmpfilename=gtmpfilename) cleans up the remote run specific directory. Patch: 7571380 cli/xml: Fix wrong XML format in volume get command broke "gluster volume get <vol> changelog.rollover-time --xml" Now fixed function utils.py::get_changelog_rollover_time() Fixed spurious trailing space getting written if second path is empty in main.py::write_output() Fixed repetitive changelog processing in changelog.py::get_changes() > Reviewed-on: http://review.gluster.org/15609 > Smoke: Gluster Build System <jenkins@build.gluster.org> > CentOS-regression: Gluster Build System <jenkins@build.gluster.org> > NetBSD-regression: NetBSD Build System <jenkins@build.gluster.org> > Reviewed-by: Aravinda VK <avishwan@redhat.com> (cherry picked from commit feea851) Change-Id: Ia8d96e2cd47bf2a64416bece312e67631a1dbf29 BUG: 1388731 Signed-off-by: Milind Changire <mchangir@redhat.com> Reviewed-on: http://review.gluster.org/15729 Smoke: Gluster Build System <jenkins@build.gluster.org> NetBSD-regression: NetBSD Build System <jenkins@build.gluster.org> CentOS-regression: Gluster Build System <jenkins@build.gluster.org> Reviewed-by: Aravinda VK <avishwan@redhat.com>

Problem: Spurious __gf_free() assertion failures seen all over the place with header->magic being overwritten when running port probing tests with 'nmap' Solution: Fix sequence of: 1. add accept()ed socket connection fd to epoll set 2. add newly created rpc_transport_t object in RPCSVC service list Correct sequence is #2 followed by #1. Reason: Adding new fd returned by accept() to epoll set causes an epoll_wait() to return immediately with a POLLIN event. This races ahead to a readv() which returms with errno:104 (Connection reset by peer) during port probing using 'nmap'. The error is then handled by POLLERR code to remove the new transport object from RPCSVC service list and later unref and destroy the rpc transport object. socket_server_event_handler() then catches up with registering the unref'd/destroyed rpc transport object. This is later manifest as assertion failures in __gf_free() with the header->magic field botched due to invalid address references. All this does not result in a Segmentation Fault since the address space continues to be mapped into the process and pages still being referenced elsewhere. As a further note: This race happens only in accept() codepath. Only in this codepath, the notify will be referring to two transports: 1, listener transport and 2. newly accepted transport All other notify refer to only one transport i.e., the transport/socket on which the event is received. Since epoll is ONE_SHOT another event won't arrive on the same socket till the current event is processed. However, in the accept() codepath, the current event - ACCEPT - and the new event - POLLIN/POLLER - arrive on two different sockets: 1. ACCEPT on listener socket and 2. POLLIN/POLLERR on newly registered socket. Also, note that these two events are handled different thread contexts. Cleanup: Critical section in socket_server_event_handler() has been removed. Instead, an additional ref on new_trans has been used to avoid ref/unref race when notifying RPCSVC. Change-Id: I4417924bc9e6277d24bd1a1c5bcb7445bcb226a3 BUG: 1438966 Signed-off-by: Milind Changire <mchangir@redhat.com> Reviewed-on: https://review.gluster.org/17139 Smoke: Gluster Build System <jenkins@build.gluster.org> NetBSD-regression: NetBSD Build System <jenkins@build.gluster.org> CentOS-regression: Gluster Build System <jenkins@build.gluster.org> Reviewed-by: Amar Tumballi <amarts@redhat.com> Reviewed-by: Oleksandr Natalenko <oleksandr@natalenko.name> Reviewed-by: Jeff Darcy <jeff@pl.atyp.us>

Problem: Spurious __gf_free() assertion failures seen all over the place with header->magic being overwritten when running port probing tests with 'nmap' Solution: Fix sequence of: 1. add accept()ed socket connection fd to epoll set 2. add newly created rpc_transport_t object in RPCSVC service list Correct sequence is #2 followed by #1. Reason: Adding new fd returned by accept() to epoll set causes an epoll_wait() to return immediately with a POLLIN event. This races ahead to a readv() which returms with errno:104 (Connection reset by peer) during port probing using 'nmap'. The error is then handled by POLLERR code to remove the new transport object from RPCSVC service list and later unref and destroy the rpc transport object. socket_server_event_handler() then catches up with registering the unref'd/destroyed rpc transport object. This is later manifest as assertion failures in __gf_free() with the header->magic field botched due to invalid address references. All this does not result in a Segmentation Fault since the address space continues to be mapped into the process and pages still being referenced elsewhere. As a further note: This race happens only in accept() codepath. Only in this codepath, the notify will be referring to two transports: 1, listener transport and 2. newly accepted transport All other notify refer to only one transport i.e., the transport/socket on which the event is received. Since epoll is ONE_SHOT another event won't arrive on the same socket till the current event is processed. However, in the accept() codepath, the current event - ACCEPT - and the new event - POLLIN/POLLER - arrive on two different sockets: 1. ACCEPT on listener socket and 2. POLLIN/POLLERR on newly registered socket. Also, note that these two events are handled different thread contexts. Cleanup: Critical section in socket_server_event_handler() has been removed. Instead, an additional ref on new_trans has been used to avoid ref/unref race when notifying RPCSVC. mainline: > BUG: 1438966 > Signed-off-by: Milind Changire <mchangir@redhat.com> > Reviewed-on: https://review.gluster.org/17139 > Smoke: Gluster Build System <jenkins@build.gluster.org> > NetBSD-regression: NetBSD Build System <jenkins@build.gluster.org> > CentOS-regression: Gluster Build System <jenkins@build.gluster.org> > Reviewed-by: Amar Tumballi <amarts@redhat.com> > Reviewed-by: Oleksandr Natalenko <oleksandr@natalenko.name> > Reviewed-by: Jeff Darcy <jeff@pl.atyp.us> (cherry picked from commit 4f7ef30) Change-Id: I4417924bc9e6277d24bd1a1c5bcb7445bcb226a3 BUG: 1449169 Signed-off-by: Milind Changire <mchangir@redhat.com> Reviewed-on: https://review.gluster.org/17217 Smoke: Gluster Build System <jenkins@build.gluster.org> NetBSD-regression: NetBSD Build System <jenkins@build.gluster.org> CentOS-regression: Gluster Build System <jenkins@build.gluster.org> Reviewed-by: Raghavendra G <rgowdapp@redhat.com>

Problem: Spurious __gf_free() assertion failures seen all over the place with header->magic being overwritten when running port probing tests with 'nmap' Solution: Fix sequence of: 1. add accept()ed socket connection fd to epoll set 2. add newly created rpc_transport_t object in RPCSVC service list Correct sequence is #2 followed by #1. Reason: Adding new fd returned by accept() to epoll set causes an epoll_wait() to return immediately with a POLLIN event. This races ahead to a readv() which returms with errno:104 (Connection reset by peer) during port probing using 'nmap'. The error is then handled by POLLERR code to remove the new transport object from RPCSVC service list and later unref and destroy the rpc transport object. socket_server_event_handler() then catches up with registering the unref'd/destroyed rpc transport object. This is later manifest as assertion failures in __gf_free() with the header->magic field botched due to invalid address references. All this does not result in a Segmentation Fault since the address space continues to be mapped into the process and pages still being referenced elsewhere. As a further note: This race happens only in accept() codepath. Only in this codepath, the notify will be referring to two transports: 1, listener transport and 2. newly accepted transport All other notify refer to only one transport i.e., the transport/socket on which the event is received. Since epoll is ONE_SHOT another event won't arrive on the same socket till the current event is processed. However, in the accept() codepath, the current event - ACCEPT - and the new event - POLLIN/POLLER - arrive on two different sockets: 1. ACCEPT on listener socket and 2. POLLIN/POLLERR on newly registered socket. Also, note that these two events are handled different thread contexts. Cleanup: Critical section in socket_server_event_handler() has been removed. Instead, an additional ref on new_trans has been used to avoid ref/unref race when notifying RPCSVC. mainline: > BUG: 1438966 > Signed-off-by: Milind Changire <mchangir@redhat.com> > Reviewed-on: https://review.gluster.org/17139 > Smoke: Gluster Build System <jenkins@build.gluster.org> > NetBSD-regression: NetBSD Build System <jenkins@build.gluster.org> > CentOS-regression: Gluster Build System <jenkins@build.gluster.org> > Reviewed-by: Amar Tumballi <amarts@redhat.com> > Reviewed-by: Oleksandr Natalenko <oleksandr@natalenko.name> > Reviewed-by: Jeff Darcy <jeff@pl.atyp.us> (cherry picked from commit 4f7ef30) Change-Id: I4417924bc9e6277d24bd1a1c5bcb7445bcb226a3 BUG: 1449191 Signed-off-by: Milind Changire <mchangir@redhat.com> Reviewed-on: https://review.gluster.org/17218 Smoke: Gluster Build System <jenkins@build.gluster.org> NetBSD-regression: NetBSD Build System <jenkins@build.gluster.org> CentOS-regression: Gluster Build System <jenkins@build.gluster.org> Reviewed-by: Shyamsundar Ranganathan <srangana@redhat.com>

…ction Problem: Sometime brick process is getting crash in notify function at the time of cleanup db connection while brick mux is enabled. Solution: In changetimerrecorder (ctr) notify function after cleanup db connection set to db_conn to NULL to avoid reuse the same db connection again. Note: Below is the backtrace pattern showing by brick process #0 0x00007ff98a30c1f7 in raise () from /lib64/libc.so.6 #1 0x00007ff98a30d8e8 in abort () from /lib64/libc.so.6 #2 0x00007ff98a34bf47 in __libc_message () from /lib64/libc.so.6 #3 0x00007ff98a351b54 in malloc_printerr () from /lib64/libc.so.6 #4 0x00007ff98a3537aa in _int_free () from /lib64/libc.so.6 #5 0x00007ff97d95e311 in gf_sql_connection_fini (sql_connection=sql_connection@entry=0x7ff8e8496b50) at gfdb_sqlite3.c:42 #6 0x00007ff97d95e38a in gf_sqlite3_fini (db_conn=0x7ff92ca04470) at gfdb_sqlite3.c:507 #7 0x00007ff97d957156 in fini_db (_conn_node=0x7ff92ca04470) at gfdb_data_store.c:326 #8 0x00007ff97db78679 in notify (this=0x7ff92c5b3670, event=9, data=0x7ff92c5b5a00) at changetimerecorder.c:2178 #9 0x00007ff98bca0dc2 in xlator_notify (xl=0x7ff92c5b3670, event=event@entry=9, data=data@entry=0x7ff92c5b5a00) at xlator.c:549 #10 0x00007ff98bd3ac12 in default_notify (this=this@entry=0x7ff92c5b5a00, event=9, data=data@entry=0x7ff92c5b6d50) at defaults.c:3139 BUG: 1475632 Change-Id: Idd4bfdb4629c4799ac477ade81228065212683fb Signed-off-by: Mohit Agrawal <moagrawa@redhat.com> Reviewed-on: https://review.gluster.org/17888 Reviewed-by: Jeff Darcy <jeff@pl.atyp.us> Smoke: Gluster Build System <jenkins@build.gluster.org> CentOS-regression: Gluster Build System <jenkins@build.gluster.org> Reviewed-by: Atin Mukherjee <amukherj@redhat.com>

…ction Problem: Sometime brick process is getting crash in notify function at the time of cleanup db connection while brick mux is enabled. Solution: In changetimerrecorder (ctr) notify function after cleanup db connection set to db_conn to NULL to avoid reuse the same db connection again. Note: Below is the backtrace pattern showing by brick process #0 0x00007ff98a30c1f7 in raise () from /lib64/libc.so.6 #1 0x00007ff98a30d8e8 in abort () from /lib64/libc.so.6 #2 0x00007ff98a34bf47 in __libc_message () from /lib64/libc.so.6 #3 0x00007ff98a351b54 in malloc_printerr () from /lib64/libc.so.6 #4 0x00007ff98a3537aa in _int_free () from /lib64/libc.so.6 #5 0x00007ff97d95e311 in gf_sql_connection_fini (sql_connection=sql_connection@entry=0x7ff8e8496b50) at gfdb_sqlite3.c:42 #6 0x00007ff97d95e38a in gf_sqlite3_fini (db_conn=0x7ff92ca04470) at gfdb_sqlite3.c:507 #7 0x00007ff97d957156 in fini_db (_conn_node=0x7ff92ca04470) at gfdb_data_store.c:326 #8 0x00007ff97db78679 in notify (this=0x7ff92c5b3670, event=9, data=0x7ff92c5b5a00) at changetimerecorder.c:2178 #9 0x00007ff98bca0dc2 in xlator_notify (xl=0x7ff92c5b3670, event=event@entry=9, data=data@entry=0x7ff92c5b5a00) at xlator.c:549 #10 0x00007ff98bd3ac12 in default_notify (this=this@entry=0x7ff92c5b5a00, event=9, data=data@entry=0x7ff92c5b6d50) at defaults.c:3139 > BUG: 1475632 > Change-Id: Idd4bfdb4629c4799ac477ade81228065212683fb > Signed-off-by: Mohit Agrawal <moagrawa@redhat.com> > Reviewed-on: https://review.gluster.org/17888 > Reviewed-by: Jeff Darcy <jeff@pl.atyp.us> > Smoke: Gluster Build System <jenkins@build.gluster.org> > CentOS-regression: Gluster Build System <jenkins@build.gluster.org> > Reviewed-by: Atin Mukherjee <amukherj@redhat.com> > (cherry picked from commit fc0fce2) BUG: 1476109 Change-Id: I96b7ab765b596cec5b779d7186ec549615e3b68b Signed-off-by: Mohit Agrawal <moagrawa@redhat.com> Reviewed-on: https://review.gluster.org/17902 Smoke: Gluster Build System <jenkins@build.gluster.org> Reviewed-by: Atin Mukherjee <amukherj@redhat.com> CentOS-regression: Gluster Build System <jenkins@build.gluster.org>

Refcounting added for nfs call state in https://review.gluster.org/17696. This is based on assumption that call state won't NULL when it is freed. But currently gluster nfs server is crashing in different scenarios at nfs3_getattr() with following bt #0 0x00007ff1cfea9205 in _gf_ref_put (ref=ref@entry=0x0) at refcount.c:36 #1 0x00007ff1c1997455 in nfs3_call_state_wipe (cs=cs@entry=0x0) at nfs3.c:559 #2 0x00007ff1c1998931 in nfs3_getattr (req=req@entry=0x7ff1bc0b26d0, fh=fh@entry=0x7ff1c2f76ae0) at nfs3.c:962 #3 0x00007ff1c1998c8a in nfs3svc_getattr (req=0x7ff1bc0b26d0) at nfs3.c:987 #4 0x00007ff1cfbfd8c5 in rpcsvc_handle_rpc_call (svc=0x7ff1bc03e500, trans=trans@entry=0x7ff1bc0c8020, msg=<optimized out>) at rpcsvc.c:695 #5 0x00007ff1cfbfdaab in rpcsvc_notify (trans=0x7ff1bc0c8020, mydata=<optimized out>, event=<optimized out>, data=<optimized out>) at rpcsvc.c:789 #6 0x00007ff1cfbff9e3 in rpc_transport_notify (this=this@entry=0x7ff1bc0c8020, event=event@entry=RPC_TRANSPORT_MSG_RECEIVED, data=data@entry=0x7ff1bc0038d0) at rpc-transport.c:538 #7 0x00007ff1c4a2e3d6 in socket_event_poll_in (this=this@entry=0x7ff1bc0c8020, notify_handled=<optimized out>) at socket.c:2306 #8 0x00007ff1c4a3097c in socket_event_handler (fd=21, idx=9, gen=19, data=0x7ff1bc0c8020, poll_in=1, poll_out=0, poll_err=0) at socket.c:2458 #9 0x00007ff1cfe950f6 in event_dispatch_epoll_handler (event=0x7ff1c2f76e80, event_pool=0x5618154d5ee0) at event-epoll.c:572 #10 event_dispatch_epoll_worker (data=0x56181551cbd0) at event-epoll.c:648 #11 0x00007ff1cec99e25 in start_thread () from /lib64/libpthread.so.0 #12 0x00007ff1ce56634d in clone () from /lib64/libc.so.6 This patch add previous NULL check move from __nfs3_call_state_wipe() to nfs3_call_state_wipe() Change-Id: I2d73632f4be23f14d8467be3d908b09b3a2d87ea BUG: 1479030 Signed-off-by: Jiffin Tony Thottan <jthottan@redhat.com> Reviewed-on: https://review.gluster.org/17989 Smoke: Gluster Build System <jenkins@build.gluster.org> CentOS-regression: Gluster Build System <jenkins@build.gluster.org> Reviewed-by: Niels de Vos <ndevos@redhat.com>

Refcounting added for nfs call state in https://review.gluster.org/17696. This is based on assumption that call state won't NULL when it is freed. But currently gluster nfs server is crashing in different scenarios at nfs3_getattr() with following bt #0 0x00007ff1cfea9205 in _gf_ref_put (ref=ref@entry=0x0) at refcount.c:36 #1 0x00007ff1c1997455 in nfs3_call_state_wipe (cs=cs@entry=0x0) at nfs3.c:559 #2 0x00007ff1c1998931 in nfs3_getattr (req=req@entry=0x7ff1bc0b26d0, fh=fh@entry=0x7ff1c2f76ae0) at nfs3.c:962 #3 0x00007ff1c1998c8a in nfs3svc_getattr (req=0x7ff1bc0b26d0) at nfs3.c:987 #4 0x00007ff1cfbfd8c5 in rpcsvc_handle_rpc_call (svc=0x7ff1bc03e500, trans=trans@entry=0x7ff1bc0c8020, msg=<optimized out>) at rpcsvc.c:695 #5 0x00007ff1cfbfdaab in rpcsvc_notify (trans=0x7ff1bc0c8020, mydata=<optimized out>, event=<optimized out>, data=<optimized out>) at rpcsvc.c:789 #6 0x00007ff1cfbff9e3 in rpc_transport_notify (this=this@entry=0x7ff1bc0c8020, event=event@entry=RPC_TRANSPORT_MSG_RECEIVED, data=data@entry=0x7ff1bc0038d0) at rpc-transport.c:538 #7 0x00007ff1c4a2e3d6 in socket_event_poll_in (this=this@entry=0x7ff1bc0c8020, notify_handled=<optimized out>) at socket.c:2306 #8 0x00007ff1c4a3097c in socket_event_handler (fd=21, idx=9, gen=19, data=0x7ff1bc0c8020, poll_in=1, poll_out=0, poll_err=0) at socket.c:2458 #9 0x00007ff1cfe950f6 in event_dispatch_epoll_handler (event=0x7ff1c2f76e80, event_pool=0x5618154d5ee0) at event-epoll.c:572 #10 event_dispatch_epoll_worker (data=0x56181551cbd0) at event-epoll.c:648 #11 0x00007ff1cec99e25 in start_thread () from /lib64/libpthread.so.0 #12 0x00007ff1ce56634d in clone () from /lib64/libc.so.6 This patch add previous NULL check move from __nfs3_call_state_wipe() to nfs3_call_state_wipe() Cherry picked from commit 111d6bd: > Change-Id: I2d73632f4be23f14d8467be3d908b09b3a2d87ea > BUG: 1479030 > Signed-off-by: Jiffin Tony Thottan <jthottan@redhat.com> > Reviewed-on: https://review.gluster.org/17989 > Smoke: Gluster Build System <jenkins@build.gluster.org> > CentOS-regression: Gluster Build System <jenkins@build.gluster.org> > Reviewed-by: Niels de Vos <ndevos@redhat.com> Change-Id: I2d73632f4be23f14d8467be3d908b09b3a2d87ea BUG: 1479263 Signed-off-by: Niels de Vos <ndevos@redhat.com> Reviewed-on: https://review.gluster.org/17994 Reviewed-by: jiffin tony Thottan <jthottan@redhat.com> Smoke: Gluster Build System <jenkins@build.gluster.org> CentOS-regression: Gluster Build System <jenkins@build.gluster.org>

Refcounting added for nfs call state in https://review.gluster.org/17696. This is based on assumption that call state won't NULL when it is freed. But currently gluster nfs server is crashing in different scenarios at nfs3_getattr() with following bt #0 0x00007ff1cfea9205 in _gf_ref_put (ref=ref@entry=0x0) at refcount.c:36 #1 0x00007ff1c1997455 in nfs3_call_state_wipe (cs=cs@entry=0x0) at nfs3.c:559 #2 0x00007ff1c1998931 in nfs3_getattr (req=req@entry=0x7ff1bc0b26d0, fh=fh@entry=0x7ff1c2f76ae0) at nfs3.c:962 #3 0x00007ff1c1998c8a in nfs3svc_getattr (req=0x7ff1bc0b26d0) at nfs3.c:987 #4 0x00007ff1cfbfd8c5 in rpcsvc_handle_rpc_call (svc=0x7ff1bc03e500, trans=trans@entry=0x7ff1bc0c8020, msg=<optimized out>) at rpcsvc.c:695 #5 0x00007ff1cfbfdaab in rpcsvc_notify (trans=0x7ff1bc0c8020, mydata=<optimized out>, event=<optimized out>, data=<optimized out>) at rpcsvc.c:789 #6 0x00007ff1cfbff9e3 in rpc_transport_notify (this=this@entry=0x7ff1bc0c8020, event=event@entry=RPC_TRANSPORT_MSG_RECEIVED, data=data@entry=0x7ff1bc0038d0) at rpc-transport.c:538 #7 0x00007ff1c4a2e3d6 in socket_event_poll_in (this=this@entry=0x7ff1bc0c8020, notify_handled=<optimized out>) at socket.c:2306 #8 0x00007ff1c4a3097c in socket_event_handler (fd=21, idx=9, gen=19, data=0x7ff1bc0c8020, poll_in=1, poll_out=0, poll_err=0) at socket.c:2458 #9 0x00007ff1cfe950f6 in event_dispatch_epoll_handler (event=0x7ff1c2f76e80, event_pool=0x5618154d5ee0) at event-epoll.c:572 #10 event_dispatch_epoll_worker (data=0x56181551cbd0) at event-epoll.c:648 #11 0x00007ff1cec99e25 in start_thread () from /lib64/libpthread.so.0 #12 0x00007ff1ce56634d in clone () from /lib64/libc.so.6 This patch add previous NULL check move from __nfs3_call_state_wipe() to nfs3_call_state_wipe() Cherry picked from commit 111d6bd: > Change-Id: I2d73632f4be23f14d8467be3d908b09b3a2d87ea > BUG: 1479030 > Signed-off-by: Jiffin Tony Thottan <jthottan@redhat.com> > Reviewed-on: https://review.gluster.org/17989 > Smoke: Gluster Build System <jenkins@build.gluster.org> > CentOS-regression: Gluster Build System <jenkins@build.gluster.org> > Reviewed-by: Niels de Vos <ndevos@redhat.com> Change-Id: I2d73632f4be23f14d8467be3d908b09b3a2d87ea BUG: 1480594 Signed-off-by: Niels de Vos <ndevos@redhat.com> Reviewed-on: https://review.gluster.org/18027 Smoke: Gluster Build System <jenkins@build.gluster.org> CentOS-regression: Gluster Build System <jenkins@build.gluster.org> Reviewed-by: Shyamsundar Ranganathan <srangana@redhat.com>

In DHT, after locks on sub-volumesare acquired, it would do rmdir in the following sequence, 1. do rmdir on all subvolumes except the hased in a loop; 2. after 1 is done, do rmdir on the hashed subvolume The problem is gluster#1 is doing a check to skip sending rmdir to hashed subvolume in the loop. If the last subvolume to check is actually the hashed one, and gluster#2 is quickly done before the loop ends, accessing shared data, which is destroyed by gluster#2, in gluster#1 would cause a crash. Fix this by moving access of shared data, that is the call frame and its local data out of the loop in gluster#1. Change-Id: I8db7cf7cb262d74efcb58eb00f02ea37df4be4e2 Signed-off-by: Zhang Huan <zhanghuan@open-fs.com>

gluster-ant · 2018-03-29T05:49:46Z

A patch https://review.gluster.org/19793 has been posted that references this issue.
Commit message: Testing a change with rfc in topic and an issue

gluster-ant · 2018-03-29T06:02:41Z

A patch https://review.gluster.org/19793 has been posted that references this issue.
Commit message: Testing a change with rfc in topic and an issue

The general idea of the changes is to prevent resetting event generation to zero in the inode ctx, since event gen is something that should follow 'causal order'. Change gluster#1: For a read txn, in inode refresh cbk, if event_generation is found zero, we are failing the read fop. This is not needed because change in event gen is only a marker for the next inode refresh to happen and should not be taken into account by the current read txn. Change gluster#2: The event gen being zero above can happen if there is a racing lookup, which resets even get (in afr_lookup_done) if there are non zero afr xattrs. The resetting is done only to trigger an inode refresh and a possible client side heal on the next lookup. That can be acheived by setting the need_refresh flag in the inode ctx. So replaced all occurences of resetting even gen to zero with a call to afr_inode_need_refresh_set(). Change gluster#3: In both lookup and discover path, we are doing an inode refresh which is not required since all 3 essentially do the same thing- update the inode ctx with the good/bad copies from the brick replies. Inode refresh also triggers background heals, but I think it is okay to do it when we call refresh during the read and write txns and not in the lookup path. The .ts which relied on inode refresh in lookup path to trigger heals are now changed to do read txn so that inode refresh and the heal happens. Upstream patch details: > Change-Id: Iebf39a9be6ffd7ffd6e4046c96b0fa78ade6c5ec > Fixes: gluster#1179 > Signed-off-by: Ravishankar N <ravishankar@redhat.com> > Reported-by: Erik Jacobson <erik.jacobson at hpe.com> Upstream patch: https://review.gluster.org/#/c/glusterfs/+/24316/ BUG: 1640148 Change-Id: Iebf39a9be6ffd7ffd6e4046c96b0fa78ade6c5ec Signed-off-by: karthik-us <ksubrahm@redhat.com> Reviewed-on: https://code.engineering.redhat.com/gerrit/222074 Tested-by: RHGS Build Bot <nigelb@redhat.com> Reviewed-by: Ravishankar Narayanankutty <ravishankar@redhat.com>

The client is throwing below stacktrace while asan is enabled. The client is facing an issue while application is trying to call removexattr in 2x1 subvol and non-mds subvol is down. As we can see in below stacktrace dht_setxattr_mds_cbk is calling dht_setxattr_non_mds_cbk and dht_setxattr_non_mds_cbk is trying to wipe local because call_cnt is 0 but dht_setxattr_mds_cbk is trying to access frame->local that;s why it is crashed. x621000051c34 is located 1844 bytes inside of 4164-byte region [0x621000051500,0x621000052544) freed by thread T7 here: #0 0x7f916ccb9388 in __interceptor_free.part.0 (/lib64/libasan.so.8+0xb9388) #1 0x7f91654af204 in dht_local_wipe /root/glusterfs_new/glusterfs/xlators/cluster/dht/src/dht-helper.c:713 #2 0x7f91654af204 in dht_setxattr_non_mds_cbk /root/glusterfs_new/glusterfs/xlators/cluster/dht/src/dht-common.c:3900 #3 0x7f91694c1f42 in client4_0_removexattr_cbk /root/glusterfs_new/glusterfs/xlators/protocol/client/src/client-rpc-fops_v2.c:1061 #4 0x7f91694ba26f in client_submit_request /root/glusterfs_new/glusterfs/xlators/protocol/client/src/client.c:288 #5 0x7f91695021bd in client4_0_removexattr /root/glusterfs_new/glusterfs/xlators/protocol/client/src/client-rpc-fops_v2.c:4480 #6 0x7f91694a5f56 in client_removexattr /root/glusterfs_new/glusterfs/xlators/protocol/client/src/client.c:1439 #7 0x7f91654a1161 in dht_setxattr_mds_cbk /root/glusterfs_new/glusterfs/xlators/cluster/dht/src/dht-common.c:3979 #8 0x7f91694c1f42 in client4_0_removexattr_cbk /root/glusterfs_new/glusterfs/xlators/protocol/client/src/client-rpc-fops_v2.c:1061 #9 0x7f916cbc4340 in rpc_clnt_handle_reply /root/glusterfs_new/glusterfs/rpc/rpc-lib/src/rpc-clnt.c:723 #10 0x7f916cbc4340 in rpc_clnt_notify /root/glusterfs_new/glusterfs/rpc/rpc-lib/src/rpc-clnt.c:890 #11 0x7f916cbb7ec5 in rpc_transport_notify /root/glusterfs_new/glusterfs/rpc/rpc-lib/src/rpc-transport.c:504 #12 0x7f916a1aa5fa in socket_event_poll_in_async /root/glusterfs_new/glusterfs/rpc/rpc-transport/socket/src/socket.c:2358 #13 0x7f916a1bd7c2 in gf_async ../../../../libglusterfs/src/glusterfs/async.h:187 #14 0x7f916a1bd7c2 in socket_event_poll_in /root/glusterfs_new/glusterfs/rpc/rpc-transport/socket/src/socket.c:2399 #15 0x7f916a1bd7c2 in socket_event_handler /root/glusterfs_new/glusterfs/rpc/rpc-transport/socket/src/socket.c:2790 #16 0x7f916a1bd7c2 in socket_event_handler /root/glusterfs_new/glusterfs/rpc/rpc-transport/socket/src/socket.c:2710 #17 0x7f916c946d22 in event_dispatch_epoll_handler /root/glusterfs_new/glusterfs/libglusterfs/src/event-epoll.c:614 #18 0x7f916c946d22 in event_dispatch_epoll_worker /root/glusterfs_new/glusterfs/libglusterfs/src/event-epoll.c:725 #19 0x7f916be8cdec in start_thread (/lib64/libc.so.6+0x8cdec) Solution: Use switch instead of using if statement to wind a operation, in case of switch the code will not try to access local after wind a operation for last dht subvol. Fixes: #3732 Change-Id: I031bc814d6df98058430ef4de7040e3370d1c677 Signed-off-by: Mohit Agrawal <moagrawa@redhat.com>

In glusterd-store.c, there are while loops like: gf_store_iter_get_next(iter, &key, &value, &op_errno); while (!ret) { if (xx_condition) { do_something(); goto out; } GF_FREE(key); GF_FREE(value); key = NULL; value = NULL; ret = gf_store_iter_get_next(iter, &key, &value, &op_errno); } It's ok in normal case, howerver, once the condition does not meet and the procedure goto 'out', there will be memory leak. Hence, it is necessary to put a check at 'out'. Similar leaks will be triggered in glusterd_store_retrieve_peers. If no peerinfo is found, the procedure will goto the next loop. It means memory allocated for key & value will be leaked once gf_store_iter_get_next is called again in the next loop. This patch fixes above mentioned memory leaks, and that detected by asan. Direct leak of 11430 byte(s) in 150 object(s) allocated from: #0 0x7f59844efbb8 in __interceptor_malloc (/lib64/libasan.so.5+0xefbb8) gluster#1 0x7f5983aeb96d in __gf_malloc (/lib64/libglusterfs.so.0+0xeb96d) gluster#2 0x7f59775b569b in glusterd_store_update_volinfo ../../../../libglusterfs/src/glusterfs/mem-pool.h:170 gluster#3 0x7f59775be3b5 in glusterd_store_retrieve_volume glusterd-store.c:3334 gluster#4 0x7f59775bf076 in glusterd_store_retrieve_volumes glusterd-store.c:3571 gluster#5 0x7f59775bfc9e in glusterd_restore glusterd-store.c:4913 gluster#6 0x7f59774ca5a1 (/usr/lib64/glusterfs/8.2/xlator/mgmt/glusterd.so+0xca5a1) gluster#7 0x7f5983a7cb6b in __xlator_init xlator.c:594 gluster#8 0x7f5983b0c5d0 in glusterfs_graph_init graph.c:422 gluster#9 0x7f5983b0d422 in glusterfs_graph_activate (/lib64/libglusterfs.so.0+0x10d422) gluster#10 0x5605f2e1eff5 in glusterfs_process_volfp glusterfsd.c:2506 gluster#11 0x5605f2e1f238 in glusterfs_volumes_init glusterfsd.c:2577 gluster#12 0x5605f2e15d8d in main (/usr/sbin/glusterfsd+0x15d8d) gluster#13 0x7f598103acf2 in __libc_start_main (/lib64/libc.so.6+0x3acf2) gluster#14 0x5605f2e162cd in _start (/usr/sbin/glusterfsd+0x162cd) Direct leak of 3351 byte(s) in 30 object(s) allocated from: #0 0x7f59844efbb8 in __interceptor_malloc (/lib64/libasan.so.5+0xefbb8) gluster#1 0x7f5983aeb96d in __gf_malloc (/lib64/libglusterfs.so.0+0xeb96d) gluster#2 0x7f59775541e7 in glusterd_is_path_in_use ../../../../libglusterfs/src/glusterfs/mem-pool.h:170 gluster#3 0x7f59775541e7 in glusterd_check_and_set_brick_xattr glusterd-utils.c:8203 gluster#4 0x7f5977554d7c in glusterd_validate_and_create_brickpath glusterd-utils.c:1549 gluster#5 0x7f5977645fcb in glusterd_op_stage_create_volume glusterd-volume-ops.c:1260 gluster#6 0x7f5977519025 in glusterd_op_stage_validate glusterd-op-sm.c:5787 gluster#7 0x7f5977672555 in gd_stage_op_phase glusterd-syncop.c:1297 gluster#8 0x7f5977676db0 in gd_sync_task_begin glusterd-syncop.c:1951 gluster#9 0x7f59776775dc in glusterd_op_begin_synctask glusterd-syncop.c:2016 gluster#10 0x7f5977642bd6 in __glusterd_handle_create_volume glusterd-volume-ops.c:506 gluster#11 0x7f59774e27b1 in glusterd_big_locked_handler glusterd-handler.c:83 gluster#12 0x7f5983b14cac in synctask_wrap syncop.c:353 gluster#13 0x7f59810240af (/lib64/libc.so.6+0x240af) Fixes: gluster#4383 Signed-off-by: chenjinhao <chen.jinhao@zte.com.cn>

Memory leaks detected by setting --enable-asan, compile, install and run gluster cmds, such as gluster v create/start/stop, etc. Direct leak of 11430 byte(s) in 150 object(s) allocated from: #0 0x7f59844efbb8 in __interceptor_malloc (/lib64/libasan.so.5+0xefbb8) gluster#1 0x7f5983aeb96d in __gf_malloc (/lib64/libglusterfs.so.0+0xeb96d) gluster#2 0x7f59775b569b in glusterd_store_update_volinfo ../../../../libglusterfs/src/glusterfs/mem-pool.h:170 gluster#3 0x7f59775be3b5 in glusterd_store_retrieve_volume glusterd-store.c:3334 gluster#4 0x7f59775bf076 in glusterd_store_retrieve_volumes glusterd-store.c:3571 gluster#5 0x7f59775bfc9e in glusterd_restore glusterd-store.c:4913 gluster#6 0x7f59774ca5a1 (/usr/lib64/glusterfs/8.2/xlator/mgmt/glusterd.so+0xca5a1) gluster#7 0x7f5983a7cb6b in __xlator_init xlator.c:594 gluster#8 0x7f5983b0c5d0 in glusterfs_graph_init graph.c:422 gluster#9 0x7f5983b0d422 in glusterfs_graph_activate (/lib64/libglusterfs.so.0+0x10d422) gluster#10 0x5605f2e1eff5 in glusterfs_process_volfp glusterfsd.c:2506 gluster#11 0x5605f2e1f238 in glusterfs_volumes_init glusterfsd.c:2577 gluster#12 0x5605f2e15d8d in main (/usr/sbin/glusterfsd+0x15d8d) gluster#13 0x7f598103acf2 in __libc_start_main (/lib64/libc.so.6+0x3acf2) gluster#14 0x5605f2e162cd in _start (/usr/sbin/glusterfsd+0x162cd) In glusterd_store_update_volinfo, memory will be leaked when the dynamic memory is put into a dict by calling dict_set_str: ret = dict_set_str(volinfo->dict, key, gf_strdup(value)); Direct leak of 3351 byte(s) in 30 object(s) allocated from: #0 0x7f59844efbb8 in __interceptor_malloc (/lib64/libasan.so.5+0xefbb8) gluster#1 0x7f5983aeb96d in __gf_malloc (/lib64/libglusterfs.so.0+0xeb96d) gluster#2 0x7f59775541e7 in glusterd_is_path_in_use ../../../../libglusterfs/src/glusterfs/mem-pool.h:170 gluster#3 0x7f59775541e7 in glusterd_check_and_set_brick_xattr glusterd-utils.c:8203 gluster#4 0x7f5977554d7c in glusterd_validate_and_create_brickpath glusterd-utils.c:1549 gluster#5 0x7f5977645fcb in glusterd_op_stage_create_volume glusterd-volume-ops.c:1260 gluster#6 0x7f5977519025 in glusterd_op_stage_validate glusterd-op-sm.c:5787 gluster#7 0x7f5977672555 in gd_stage_op_phase glusterd-syncop.c:1297 gluster#8 0x7f5977676db0 in gd_sync_task_begin glusterd-syncop.c:1951 gluster#9 0x7f59776775dc in glusterd_op_begin_synctask glusterd-syncop.c:2016 gluster#10 0x7f5977642bd6 in __glusterd_handle_create_volume glusterd-volume-ops.c:506 gluster#11 0x7f59774e27b1 in glusterd_big_locked_handler glusterd-handler.c:83 gluster#12 0x7f5983b14cac in synctask_wrap syncop.c:353 gluster#13 0x7f59810240af (/lib64/libc.so.6+0x240af) During volume creation, glusterd_is_path_in_use will be called to check brick path reuse. Under normal circumstances, there is no problem， however, when a `force` cmd is given during volume creation and the futher sys_lsetxattr failed, the memory area previously pointed by *op_errstr will be leakd, cause: out: if (strlen(msg)) *op_errstr = gf_strdup(msg); Similar leak also exists in posix_cs_set_state: value = GF_CALLOC(1, xattrsize + 1, gf_posix_mt_char); ... dict_set_str_sizen(*rsp, GF_CS_OBJECT_REMOTE, value); Signed-off-by: chenjinhao <chen.jinhao@zte.com.cn>

* glusterd: fix memory leaks detected by asan Memory leaks detected by setting --enable-asan, compile, install and run gluster cmds, such as gluster v create/start/stop, etc. Direct leak of 11430 byte(s) in 150 object(s) allocated from: #0 0x7f59844efbb8 in __interceptor_malloc (/lib64/libasan.so.5+0xefbb8) #1 0x7f5983aeb96d in __gf_malloc (/lib64/libglusterfs.so.0+0xeb96d) #2 0x7f59775b569b in glusterd_store_update_volinfo ../../../../libglusterfs/src/glusterfs/mem-pool.h:170 #3 0x7f59775be3b5 in glusterd_store_retrieve_volume glusterd-store.c:3334 #4 0x7f59775bf076 in glusterd_store_retrieve_volumes glusterd-store.c:3571 #5 0x7f59775bfc9e in glusterd_restore glusterd-store.c:4913 #6 0x7f59774ca5a1 (/usr/lib64/glusterfs/8.2/xlator/mgmt/glusterd.so+0xca5a1) #7 0x7f5983a7cb6b in __xlator_init xlator.c:594 #8 0x7f5983b0c5d0 in glusterfs_graph_init graph.c:422 #9 0x7f5983b0d422 in glusterfs_graph_activate (/lib64/libglusterfs.so.0+0x10d422) #10 0x5605f2e1eff5 in glusterfs_process_volfp glusterfsd.c:2506 #11 0x5605f2e1f238 in glusterfs_volumes_init glusterfsd.c:2577 #12 0x5605f2e15d8d in main (/usr/sbin/glusterfsd+0x15d8d) #13 0x7f598103acf2 in __libc_start_main (/lib64/libc.so.6+0x3acf2) #14 0x5605f2e162cd in _start (/usr/sbin/glusterfsd+0x162cd) In glusterd_store_update_volinfo, memory will be leaked when the dynamic memory is put into a dict by calling dict_set_str: ret = dict_set_str(volinfo->dict, key, gf_strdup(value)); Direct leak of 3351 byte(s) in 30 object(s) allocated from: #0 0x7f59844efbb8 in __interceptor_malloc (/lib64/libasan.so.5+0xefbb8) #1 0x7f5983aeb96d in __gf_malloc (/lib64/libglusterfs.so.0+0xeb96d) #2 0x7f59775541e7 in glusterd_is_path_in_use ../../../../libglusterfs/src/glusterfs/mem-pool.h:170 #3 0x7f59775541e7 in glusterd_check_and_set_brick_xattr glusterd-utils.c:8203 #4 0x7f5977554d7c in glusterd_validate_and_create_brickpath glusterd-utils.c:1549 #5 0x7f5977645fcb in glusterd_op_stage_create_volume glusterd-volume-ops.c:1260 #6 0x7f5977519025 in glusterd_op_stage_validate glusterd-op-sm.c:5787 #7 0x7f5977672555 in gd_stage_op_phase glusterd-syncop.c:1297 #8 0x7f5977676db0 in gd_sync_task_begin glusterd-syncop.c:1951 #9 0x7f59776775dc in glusterd_op_begin_synctask glusterd-syncop.c:2016 #10 0x7f5977642bd6 in __glusterd_handle_create_volume glusterd-volume-ops.c:506 #11 0x7f59774e27b1 in glusterd_big_locked_handler glusterd-handler.c:83 #12 0x7f5983b14cac in synctask_wrap syncop.c:353 #13 0x7f59810240af (/lib64/libc.so.6+0x240af) During volume creation, glusterd_is_path_in_use will be called to check brick path reuse. Under normal circumstances, there is no problem， however, when a `force` cmd is given during volume creation and the futher sys_lsetxattr failed, the memory area previously pointed by *op_errstr will be leakd, cause: out: if (strlen(msg)) *op_errstr = gf_strdup(msg); Similar leak also exists in posix_cs_set_state: value = GF_CALLOC(1, xattrsize + 1, gf_posix_mt_char); ... dict_set_str_sizen(*rsp, GF_CS_OBJECT_REMOTE, value); Signed-off-by: chenjinhao <chen.jinhao@zte.com.cn> * glusterd: fix memory leaks due to lack of GF_FREE In glusterd-store.c, there are while loops like: gf_store_iter_get_next(iter, &key, &value, &op_errno); while (!ret) { if (xx_condition) { do_something(); goto out; } GF_FREE(key); GF_FREE(value); key = NULL; value = NULL; ret = gf_store_iter_get_next(iter, &key, &value, &op_errno); } It's ok under normal case, howerver, once the condition does not meet and the procedure goto 'out', there will be memory leak. Hence, it is necessary to put a check at 'out'. Similar leaks will be triggered in glusterd_store_retrieve_peers. If no peerinfo is found, the procedure will goto the next loop. It means memory previously allocated for key & value will be leaked once gf_store_iter_get_next is called again in the next loop. Signed-off-by: chenjinhao <chen.jinhao@zte.com.cn> --------- Signed-off-by: chenjinhao <chen.jinhao@zte.com.cn> Co-authored-by: chenjinhao <chen.jinhao@zte.com.cn>

* glusterd: fix memory leaks detected by asan Memory leaks detected by setting --enable-asan, compile, install and run gluster cmds, such as gluster v create/start/stop, etc. Direct leak of 11430 byte(s) in 150 object(s) allocated from: #0 0x7f59844efbb8 in __interceptor_malloc (/lib64/libasan.so.5+0xefbb8) gluster#1 0x7f5983aeb96d in __gf_malloc (/lib64/libglusterfs.so.0+0xeb96d) gluster#2 0x7f59775b569b in glusterd_store_update_volinfo ../../../../libglusterfs/src/glusterfs/mem-pool.h:170 gluster#3 0x7f59775be3b5 in glusterd_store_retrieve_volume glusterd-store.c:3334 gluster#4 0x7f59775bf076 in glusterd_store_retrieve_volumes glusterd-store.c:3571 gluster#5 0x7f59775bfc9e in glusterd_restore glusterd-store.c:4913 gluster#6 0x7f59774ca5a1 (/usr/lib64/glusterfs/8.2/xlator/mgmt/glusterd.so+0xca5a1) gluster#7 0x7f5983a7cb6b in __xlator_init xlator.c:594 gluster#8 0x7f5983b0c5d0 in glusterfs_graph_init graph.c:422 gluster#9 0x7f5983b0d422 in glusterfs_graph_activate (/lib64/libglusterfs.so.0+0x10d422) gluster#10 0x5605f2e1eff5 in glusterfs_process_volfp glusterfsd.c:2506 gluster#11 0x5605f2e1f238 in glusterfs_volumes_init glusterfsd.c:2577 gluster#12 0x5605f2e15d8d in main (/usr/sbin/glusterfsd+0x15d8d) gluster#13 0x7f598103acf2 in __libc_start_main (/lib64/libc.so.6+0x3acf2) gluster#14 0x5605f2e162cd in _start (/usr/sbin/glusterfsd+0x162cd) In glusterd_store_update_volinfo, memory will be leaked when the dynamic memory is put into a dict by calling dict_set_str: ret = dict_set_str(volinfo->dict, key, gf_strdup(value)); Direct leak of 3351 byte(s) in 30 object(s) allocated from: #0 0x7f59844efbb8 in __interceptor_malloc (/lib64/libasan.so.5+0xefbb8) gluster#1 0x7f5983aeb96d in __gf_malloc (/lib64/libglusterfs.so.0+0xeb96d) gluster#2 0x7f59775541e7 in glusterd_is_path_in_use ../../../../libglusterfs/src/glusterfs/mem-pool.h:170 gluster#3 0x7f59775541e7 in glusterd_check_and_set_brick_xattr glusterd-utils.c:8203 gluster#4 0x7f5977554d7c in glusterd_validate_and_create_brickpath glusterd-utils.c:1549 gluster#5 0x7f5977645fcb in glusterd_op_stage_create_volume glusterd-volume-ops.c:1260 gluster#6 0x7f5977519025 in glusterd_op_stage_validate glusterd-op-sm.c:5787 gluster#7 0x7f5977672555 in gd_stage_op_phase glusterd-syncop.c:1297 gluster#8 0x7f5977676db0 in gd_sync_task_begin glusterd-syncop.c:1951 gluster#9 0x7f59776775dc in glusterd_op_begin_synctask glusterd-syncop.c:2016 gluster#10 0x7f5977642bd6 in __glusterd_handle_create_volume glusterd-volume-ops.c:506 gluster#11 0x7f59774e27b1 in glusterd_big_locked_handler glusterd-handler.c:83 gluster#12 0x7f5983b14cac in synctask_wrap syncop.c:353 gluster#13 0x7f59810240af (/lib64/libc.so.6+0x240af) During volume creation, glusterd_is_path_in_use will be called to check brick path reuse. Under normal circumstances, there is no problem， however, when a `force` cmd is given during volume creation and the futher sys_lsetxattr failed, the memory area previously pointed by *op_errstr will be leakd, cause: out: if (strlen(msg)) *op_errstr = gf_strdup(msg); Similar leak also exists in posix_cs_set_state: value = GF_CALLOC(1, xattrsize + 1, gf_posix_mt_char); ... dict_set_str_sizen(*rsp, GF_CS_OBJECT_REMOTE, value); Signed-off-by: chenjinhao <chen.jinhao@zte.com.cn> * glusterd: fix memory leaks due to lack of GF_FREE In glusterd-store.c, there are while loops like: gf_store_iter_get_next(iter, &key, &value, &op_errno); while (!ret) { if (xx_condition) { do_something(); goto out; } GF_FREE(key); GF_FREE(value); key = NULL; value = NULL; ret = gf_store_iter_get_next(iter, &key, &value, &op_errno); } It's ok under normal case, howerver, once the condition does not meet and the procedure goto 'out', there will be memory leak. Hence, it is necessary to put a check at 'out'. Similar leaks will be triggered in glusterd_store_retrieve_peers. If no peerinfo is found, the procedure will goto the next loop. It means memory previously allocated for key & value will be leaked once gf_store_iter_get_next is called again in the next loop. Signed-off-by: chenjinhao <chen.jinhao@zte.com.cn> --------- Signed-off-by: chenjinhao <chen.jinhao@zte.com.cn> Co-authored-by: chenjinhao <chen.jinhao@zte.com.cn>

asan reports: Direct leak of 4386 byte(s) in 34 object(s) allocated from: #0 0x7ff3b72efbb8 in __interceptor_malloc (/lib64/libasan.so.5+0xefbb8) gluster#1 0x7ff3b6902420 in __gf_malloc (/lib64/libglusterfs.so.0+0x102420) gluster#2 0x7ff3b6876e18 in dict_allocate_and_serialize (/lib64/libglusterfs.so.0+0x76e18) gluster#3 0x7ff3aa5ec701 in glusterd_mgmt_handshake glusterd-handshake.c:2212 gluster#4 0x7ff3aa5ee2cc in __glusterd_peer_dump_version_cbk glusterd-handshake.c:2388 gluster#5 0x7ff3aa5b58ca in glusterd_big_locked_cbk glusterd-rpc-ops.c:217 gluster#6 0x7ff3b642ffcb in rpc_clnt_handle_reply rpc-clnt.c:780 gluster#7 0x7ff3b64307e9 in rpc_clnt_notify rpc-clnt.c:957 gluster#8 0x7ff3b6425c2c in rpc_transport_notify (/lib64/libgfrpc.so.0+0x25c2c) gluster#9 0x7ff3a880d874 in socket_event_poll_in_async socket.c:2531 gluster#10 0x7ff3a882828e in socket_event_poll_in ../../../../libglusterfs/src/glusterfs/async.h:189 gluster#11 0x7ff3a882828e in socket_event_handler socket.c:2963 gluster#12 0x7ff3a882828e in socket_event_handler socket.c:2883 gluster#13 0x7ff3b69af776 in event_dispatch_epoll_handler event-epoll.c:640 gluster#14 0x7ff3b69af776 in event_dispatch_epoll_worker event-epoll.c:751 gluster#15 0x7ff3b48082fe in start_thread pthread_create.c:479 gluster#16 0x7ff3b3e39dd2 in __GI___clone (/lib64/libc.so.6+0x39dd2) Fixes: gluster#4558 Signed-off-by: chenjinhao <chen.jinhao@zte.com.cn>

avati merged this pull request into gluster:master Apr 25, 2011

vpshastry pushed a commit to vpshastry/glusterfs that referenced this pull request Jun 30, 2014

Merge pull request gluster#1 from vpshastry/better-peer-identification

3a5bfa1

glusterd: Add hostname to list instead of replaceing upon update

rootfs pushed a commit to rootfs/glusterfs that referenced this pull request Mar 3, 2015

Merge pull request gluster#1 from gluster/master

12a7530

Resync

als-git mentioned this pull request Dec 28, 2022

glusterfs client crashes with Bus error on mount on sparc64 #3934

Closed

shenwen2012 mentioned this pull request Jan 11, 2023

version 7.9 add-brick fuse client process crash #3953

Open

feizai131 mentioned this pull request Jan 13, 2023

readdir core，access to freed memory #3955

Closed

JamesWSWu mentioned this pull request Mar 30, 2023

In debug mode, the fuse process crashes in GF_ ASSERT(!(*take_lock)) #4091

Closed

handrea2009 mentioned this pull request Apr 11, 2023

Quotad is crashing #4095

Open

SDPervez mentioned this pull request May 5, 2023

glusterfs Too many open files - Need a way to increase open files limit for bricks process glusterfsd #4134

Open

Tom-python0121 mentioned this pull request May 17, 2023

test glusterfs@11.0 failed on openEuler20_aarch64 #4151

Open

rhymerjr mentioned this pull request May 19, 2023

inode resource leak - v11.0 (maybe related to "No space left on device") #4157

Closed

amurashkin17 mentioned this pull request May 21, 2023

glusterfs 11 - SIGABRT in gf_io_uring_cq_process_some (libglusterfs.so.0+0xb49c5) #4159

Open

mohit84 mentioned this pull request Jul 5, 2023

core: glusterfs encounter a SIGSEGV in __gf_free #4190

Closed

This was referenced Oct 16, 2023

glfsheal coredump occasionally #4239

Closed

core: glfsheal encounter a SIGSEGV in __strftime_internal #4240

Merged

mjolnirgs mentioned this pull request Oct 20, 2023

glusterfsd coredumps frequently on 10.4 cluster in tcmalloc::ThreadCache::ReleaseToCentralCache #4241

Closed

mohit84 mentioned this pull request Oct 20, 2023

dht: fix asan use-after-free bug #4242

Merged

SowjanyaKotha mentioned this pull request Nov 27, 2023

Segmentation fault in gluster client #4271

Open

handrea2009 mentioned this pull request Dec 15, 2023

No documented procedure to replace a failed brick when brick path doesn't change #4286

Closed

powerfooI mentioned this pull request Mar 5, 2024

fallocate with flag FALLOC_FL_ZERO_RANGE failed #4313

Closed

wonderful1984 mentioned this pull request Mar 25, 2024

Segmentation fault gluster volume status all clients #4323

Open

l392zhan mentioned this pull request Jul 9, 2024

Memory Leaks Detected #4391

Closed

eeraser710 mentioned this pull request Dec 4, 2024

Frequent "Transport endpoint is not connected" across nodes #4435

Closed

feizai131 mentioned this pull request Jan 1, 2025

__socket_cached_read opvector的 len 非法，导致req_len 0，导致返回0 认为对端关闭socket 反复重启链接 #4448

Open

YifanZhang-311 mentioned this pull request Dec 23, 2025

Null pointer dereference in afr_notify() due to access after free in afr_has_quorum #4644

Open

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Server-enforced ACLs#1

Server-enforced ACLs#1
avati merged 1 commit intogluster:masterfrom
jdarcy:master

jdarcy commented Apr 19, 2011

Uh oh!

abperiasamy commented Apr 19, 2011

Uh oh!

avati commented Apr 25, 2011

Uh oh!

jdarcy commented Apr 25, 2011

Uh oh!

abperiasamy commented Apr 25, 2011

Uh oh!

avati commented Apr 25, 2011

Uh oh!

abperiasamy commented Apr 25, 2011

Uh oh!

jdarcy commented Apr 26, 2011

Uh oh!

avati commented Apr 26, 2011

Uh oh!

gluster-ant commented Mar 29, 2018

Uh oh!

gluster-ant commented Mar 29, 2018

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

4 participants

Conversation

jdarcy commented Apr 19, 2011

Uh oh!

abperiasamy commented Apr 19, 2011

Uh oh!

avati commented Apr 25, 2011

Uh oh!

jdarcy commented Apr 25, 2011

Uh oh!

abperiasamy commented Apr 25, 2011

Uh oh!

avati commented Apr 25, 2011

Uh oh!

abperiasamy commented Apr 25, 2011

Uh oh!

jdarcy commented Apr 26, 2011

Uh oh!

avati commented Apr 26, 2011

Uh oh!

gluster-ant commented Mar 29, 2018

Uh oh!

gluster-ant commented Mar 29, 2018

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

4 participants