Note: This bug is displayed in read-only format because the product is no longer active in Red Hat Bugzilla.
RHEL Engineering is moving the tracking of its product development work on RHEL 6 through RHEL 9 to Red Hat Jira (issues.redhat.com). If you're a Red Hat customer, please continue to file support cases via the Red Hat customer portal. If you're not, please head to the "RHEL project" in Red Hat Jira and file new tickets here. Individual Bugzilla bugs in the statuses "NEW", "ASSIGNED", and "POST" are being migrated throughout September 2023. Bugs of Red Hat partners with an assigned Engineering Partner Manager (EPM) are migrated in late September as per pre-agreed dates. Bugs against components "kernel", "kernel-rt", and "kpatch" are only migrated if still in "NEW" or "ASSIGNED". If you cannot log in to RH Jira, please consult article #7032570. That failing, please send an e-mail to the RH Jira admins at rh-issues@redhat.com to troubleshoot your issue as a user management inquiry. The email creates a ServiceNow ticket with Red Hat. Individual Bugzilla bugs that are migrated will be moved to status "CLOSED", resolution "MIGRATED", and set with "MigratedToJIRA" in "Keywords". The link to the successor Jira issue will be found under "Links", have a little "two-footprint" icon next to it, and direct you to the "RHEL project" in Red Hat Jira (issue links are of type "https://issues.redhat.com/browse/RHEL-XXXX", where "X" is a digit). This same link will be available in a blue banner at the top of the page informing you that that bug has been migrated.

Bug 1247846

Summary: epoll_ctl(del): Bad file descriptor (9)
Product: Red Hat Enterprise Linux 7 Reporter: Sam McLeod <mailinglists>
Component: libqbAssignee: Christine Caulfield <ccaulfie>
Status: CLOSED ERRATA QA Contact: cluster-qe <cluster-qe>
Severity: unspecified Docs Contact:
Priority: unspecified    
Version: 7.3CC: ccaulfie, cluster-maint, jkortus, kgaillot, mailinglists, mkelly
Target Milestone: rc   
Target Release: ---   
Hardware: x86_64   
OS: Linux   
Whiteboard:
Fixed In Version: Doc Type: Bug Fix
Doc Text:
Story Points: ---
Clone Of: Environment:
Last Closed: 2016-11-04 04:40:27 UTC Type: Bug
Regression: --- Mount Type: ---
Documentation: --- CRM:
Verified Versions: Category: ---
oVirt Team: --- RHEL 7.3 requirements from Atomic Host:
Cloudforms Team: --- Target Upstream Version:
Embargoed:
Attachments:
Description Flags
corosync config and logs none

Description Sam McLeod 2015-07-29 05:00:50 UTC 
Created attachment 1057213 [details]
corosync config and logs

Description of problem:

Corosync logs errors: epoll_ctl(del): Bad file descriptor (9) when trying to unallocated shared memory when closing ringbuffer

Example:

739467:Jul 29 14:09:52 [12727] s1-san6   crm_node:    debug: qb_rb_open_2: 	shm size:1048589; real_size:1052672; rb->word_size:263168
739468:Jul 29 14:09:52 [1379] s1-san6 corosync debug   [QB    ] qb_ipcs_disconnect(1500-12725-26) state:2
739469:Jul 29 14:09:52 [12725] s1-san6   crm_node:    debug: qb_rb_close: 	Closing ringbuffer: /dev/shm/qb-cpg-response-1500-12725-26-header
739470:Jul 29 14:09:52 [1379] s1-san6 corosync debug   [QB    ] epoll_ctl(del): Bad file descriptor (9)
739471:Jul 29 14:09:52 [1379] s1-san6 corosync debug   [MAIN  ] cs_ipcs_connection_closed()
739472:Jul 29 14:09:52 [1379] s1-san6 corosync debug   [CPG   ] exit_fn for conn=0x5654f588a0c0
739473:Jul 29 14:09:52 [12725] s1-san6   crm_node:    debug: qb_rb_close: 	Closing ringbuffer: /dev/shm/qb-cpg-event-1500-12725-26-header
739474:Jul 29 14:09:52 [1379] s1-san6 corosync debug   [MAIN  ] cs_ipcs_connection_destroyed()


Version-Release number of selected component (if applicable):

- Pacemaker 1.1.12
- Corosync 2.3.4
- CentOS Linux release 7.1.1503 (Core)
- STONITH disabled for debugging purposes
- System is not in production (yet)

How reproducible:

- Can reproduce on all four nodes on two separate clusters.
- Already occurred 12,240 times today:

root@s1-san6:~  # zgrep -ni "descriptor" /var/log/cluster/corosync.log|wc -l
12240
Comment 1 Sam McLeod 2015-07-29 05:04:37 UTC 
Interestingly some of the files in /dev/shm are owned by root and others by hacluster - not sure if this is expected or not.

root@s1-san6:/dev/shm  # ls -lah
total 39M
drwxrwxrwt  2 root      root     1.9K Jul 29 15:03 .
drwxr-xr-x 23 root      root     3.9K Jul 29 09:28 ..
-rw-rw----  1 hacluster haclient 132K Jul 29 09:11 qb-attrd-event-2378-2381-9-data
-rw-rw----  1 hacluster haclient 8.1K Jul 29 09:11 qb-attrd-event-2378-2381-9-header
-rw-rw----  1 hacluster haclient 132K Jul 29 09:11 qb-attrd-request-2378-2381-9-data
-rw-rw----  1 hacluster haclient 8.1K Jul 29 09:11 qb-attrd-request-2378-2381-9-header
-rw-rw----  1 hacluster haclient 132K Jul 29 09:11 qb-attrd-response-2378-2381-9-data
-rw-rw----  1 hacluster haclient 8.1K Jul 29 09:11 qb-attrd-response-2378-2381-9-header
-rw-------  1 root      root     1.1M Jul 29 09:11 qb-cfg-event-1500-2281-18-data
-rw-------  1 root      root     8.1K Jul 29 09:11 qb-cfg-event-1500-2281-18-header
-rw-------  1 root      root     1.1M Jul 29 09:11 qb-cfg-request-1500-2281-18-data
-rw-------  1 root      root     8.1K Jul 29 09:11 qb-cfg-request-1500-2281-18-header
-rw-------  1 root      root     1.1M Jul 29 09:11 qb-cfg-response-1500-2281-18-data
-rw-------  1 root      root     8.1K Jul 29 09:11 qb-cfg-response-1500-2281-18-header
-rw-------  1 hacluster haclient 516K Jul 29 09:11 qb-cib_rw-event-2372-2373-11-data
-rw-------  1 hacluster haclient 8.1K Jul 29 09:11 qb-cib_rw-event-2372-2373-11-header
-rw-rw----  1 hacluster haclient 516K Jul 29 09:11 qb-cib_rw-event-2372-2378-12-data
-rw-rw----  1 hacluster haclient 8.1K Jul 29 09:11 qb-cib_rw-event-2372-2378-12-header
-rw-------  1 hacluster haclient 516K Jul 29 09:11 qb-cib_rw-request-2372-2373-11-data
-rw-------  1 hacluster haclient 8.1K Jul 29 09:11 qb-cib_rw-request-2372-2373-11-header
-rw-rw----  1 hacluster haclient 516K Jul 29 09:11 qb-cib_rw-request-2372-2378-12-data
-rw-rw----  1 hacluster haclient 8.1K Jul 29 09:11 qb-cib_rw-request-2372-2378-12-header
-rw-------  1 hacluster haclient 516K Jul 29 09:11 qb-cib_rw-response-2372-2373-11-data
-rw-------  1 hacluster haclient 8.1K Jul 29 09:11 qb-cib_rw-response-2372-2373-11-header
-rw-rw----  1 hacluster haclient 516K Jul 29 09:11 qb-cib_rw-response-2372-2378-12-data
-rw-rw----  1 hacluster haclient 8.1K Jul 29 09:11 qb-cib_rw-response-2372-2378-12-header
-rw-rw----  1 hacluster haclient 516K Jul 29 09:11 qb-cib_shm-event-2372-2381-10-data
-rw-rw----  1 hacluster haclient 8.1K Jul 29 09:11 qb-cib_shm-event-2372-2381-10-header
-rw-rw----  1 hacluster haclient 516K Jul 29 09:51 qb-cib_shm-request-2372-2381-10-data
-rw-rw----  1 hacluster haclient 8.1K Jul 29 09:11 qb-cib_shm-request-2372-2381-10-header
-rw-rw----  1 hacluster haclient 516K Jul 29 09:11 qb-cib_shm-response-2372-2381-10-data
-rw-rw----  1 hacluster haclient 8.1K Jul 29 09:11 qb-cib_shm-response-2372-2381-10-header
-rw-------  1 root      root     8.1M Jul 29 11:02 qb-corosync-1379-blackbox-data
-rw-------  1 root      root     8.1K Jul 29 09:11 qb-corosync-1379-blackbox-header
-rw-------  1 root      root     1.1M Jul 29 09:11 qb-cpg-event-1500-2281-19-data
-rw-------  1 root      root     8.1K Jul 29 09:11 qb-cpg-event-1500-2281-19-header
-rw-------  1 hacluster haclient 1.1M Jul 29 09:11 qb-cpg-event-1500-2372-25-data
-rw-------  1 hacluster haclient 8.1K Jul 29 09:11 qb-cpg-event-1500-2372-25-header
-rw-------  1 root      root     1.1M Jul 29 09:11 qb-cpg-event-1500-2373-21-data
-rw-------  1 root      root     8.1K Jul 29 09:11 qb-cpg-event-1500-2373-21-header
-rw-------  1 hacluster haclient 1.1M Jul 29 09:11 qb-cpg-event-1500-2378-23-data
-rw-------  1 hacluster haclient 8.1K Jul 29 09:11 qb-cpg-event-1500-2378-23-header
-rw-------  1 hacluster haclient 1.1M Jul 29 09:11 qb-cpg-event-1500-2381-22-data
-rw-------  1 hacluster haclient 8.1K Jul 29 09:11 qb-cpg-event-1500-2381-22-header
-rw-------  1 root      root     1.1M Jul 29 09:11 qb-cpg-request-1500-2281-19-data
-rw-------  1 root      root     8.1K Jul 29 09:11 qb-cpg-request-1500-2281-19-header
-rw-------  1 hacluster haclient 1.1M Jul 29 09:51 qb-cpg-request-1500-2372-25-data
-rw-------  1 hacluster haclient 8.1K Jul 29 09:11 qb-cpg-request-1500-2372-25-header
-rw-------  1 root      root     1.1M Jul 29 09:11 qb-cpg-request-1500-2373-21-data
-rw-------  1 root      root     8.1K Jul 29 09:11 qb-cpg-request-1500-2373-21-header
-rw-------  1 hacluster haclient 1.1M Jul 29 10:26 qb-cpg-request-1500-2378-23-data
-rw-------  1 hacluster haclient 8.1K Jul 29 09:11 qb-cpg-request-1500-2378-23-header
-rw-------  1 hacluster haclient 1.1M Jul 29 09:51 qb-cpg-request-1500-2381-22-data
-rw-------  1 hacluster haclient 8.1K Jul 29 09:11 qb-cpg-request-1500-2381-22-header
-rw-------  1 root      root     1.1M Jul 29 09:11 qb-cpg-response-1500-2281-19-data
-rw-------  1 root      root     8.1K Jul 29 09:11 qb-cpg-response-1500-2281-19-header
-rw-------  1 hacluster haclient 1.1M Jul 29 09:11 qb-cpg-response-1500-2372-25-data
-rw-------  1 hacluster haclient 8.1K Jul 29 09:11 qb-cpg-response-1500-2372-25-header
-rw-------  1 root      root     1.1M Jul 29 09:11 qb-cpg-response-1500-2373-21-data
-rw-------  1 root      root     8.1K Jul 29 09:11 qb-cpg-response-1500-2373-21-header
-rw-------  1 hacluster haclient 1.1M Jul 29 09:11 qb-cpg-response-1500-2378-23-data
-rw-------  1 hacluster haclient 8.1K Jul 29 09:11 qb-cpg-response-1500-2378-23-header
-rw-------  1 hacluster haclient 1.1M Jul 29 09:11 qb-cpg-response-1500-2381-22-data
-rw-------  1 hacluster haclient 8.1K Jul 29 09:11 qb-cpg-response-1500-2381-22-header
-rw-rw----  1 hacluster haclient 132K Jul 29 09:11 qb-lrmd-event-2377-2381-7-data
-rw-rw----  1 hacluster haclient 8.1K Jul 29 09:11 qb-lrmd-event-2377-2381-7-header
-rw-rw----  1 hacluster haclient 132K Jul 29 09:51 qb-lrmd-request-2377-2381-7-data
-rw-rw----  1 hacluster haclient 8.1K Jul 29 09:11 qb-lrmd-request-2377-2381-7-header
-rw-rw----  1 hacluster haclient 132K Jul 29 09:11 qb-lrmd-response-2377-2381-7-data
-rw-rw----  1 hacluster haclient 8.1K Jul 29 09:11 qb-lrmd-response-2377-2381-7-header
-rw-------  1 root      root     1.1M Jul 29 09:11 qb-quorum-event-1500-2281-20-data
-rw-------  1 root      root     8.1K Jul 29 09:11 qb-quorum-event-1500-2281-20-header
-rw-------  1 hacluster haclient 1.1M Jul 29 09:11 qb-quorum-event-1500-2381-24-data
-rw-------  1 hacluster haclient 8.1K Jul 29 09:11 qb-quorum-event-1500-2381-24-header
-rw-------  1 root      root     1.1M Jul 29 09:11 qb-quorum-request-1500-2281-20-data
-rw-------  1 root      root     8.1K Jul 29 09:11 qb-quorum-request-1500-2281-20-header
-rw-------  1 hacluster haclient 1.1M Jul 29 09:11 qb-quorum-request-1500-2381-24-data
-rw-------  1 hacluster haclient 8.1K Jul 29 09:11 qb-quorum-request-1500-2381-24-header
-rw-------  1 root      root     1.1M Jul 29 09:11 qb-quorum-response-1500-2281-20-data
-rw-------  1 root      root     8.1K Jul 29 09:11 qb-quorum-response-1500-2281-20-header
-rw-------  1 hacluster haclient 1.1M Jul 29 09:11 qb-quorum-response-1500-2381-24-data
-rw-------  1 hacluster haclient 8.1K Jul 29 09:11 qb-quorum-response-1500-2381-24-header
-rw-------  1 root      root     132K Jul 29 09:11 qb-stonith-ng-event-2373-2377-10-data
-rw-------  1 root      root     8.1K Jul 29 09:11 qb-stonith-ng-event-2373-2377-10-header
-rw-rw----  1 hacluster haclient 132K Jul 29 09:11 qb-stonith-ng-event-2373-2381-9-data
-rw-rw----  1 hacluster haclient 8.1K Jul 29 09:11 qb-stonith-ng-event-2373-2381-9-header
-rw-------  1 root      root     132K Jul 29 09:11 qb-stonith-ng-request-2373-2377-10-data
-rw-------  1 root      root     8.1K Jul 29 09:11 qb-stonith-ng-request-2373-2377-10-header
-rw-rw----  1 hacluster haclient 132K Jul 29 09:11 qb-stonith-ng-request-2373-2381-9-data
-rw-rw----  1 hacluster haclient 8.1K Jul 29 09:11 qb-stonith-ng-request-2373-2381-9-header
-rw-------  1 root      root     132K Jul 29 09:11 qb-stonith-ng-response-2373-2377-10-data
-rw-------  1 root      root     8.1K Jul 29 09:11 qb-stonith-ng-response-2373-2377-10-header
-rw-rw----  1 hacluster haclient 132K Jul 29 09:11 qb-stonith-ng-response-2373-2381-9-data
-rw-rw----  1 hacluster haclient 8.1K Jul 29 09:11 qb-stonith-ng-response-2373-2381-9-header
Comment 3 Jan Friesse 2015-07-30 07:19:21 UTC 
Different owners (hacluster vs root) are expected and it's not a problem (file is created by client, so if client is running as hacluster, file is owned by hacluster).

epollctl is also not a big deal, but may be confusing.

Sadly corosync has no way to handle this error messages, so reassigning to libqb developers to propose solution.
Comment 5 Christine Caulfield 2015-09-24 13:29:23 UTC 
This certainly seems to be a libqb bug. libqb closes the fd before removing it from the epoll list.

This is a patch I'm testing. It just swaps over the close and epoll_del:

diff --git a/lib/ipc_shm.c b/lib/ipc_shm.c
index 5eaf245..a038d8d 100644
--- a/lib/ipc_shm.c
+++ b/lib/ipc_shm.c
@@ -228,8 +228,8 @@ qb_ipcs_shm_disconnect(struct qb_ipcs_connection *c)
        if (c->state == QB_IPCS_CONNECTION_ESTABLISHED ||
            c->state == QB_IPCS_CONNECTION_ACTIVE) {
                if (c->setup.u.us.sock > 0) {
-                       qb_ipcc_us_sock_close(c->setup.u.us.sock);
                        (void)c->service->poll_fns.dispatch_del(c->setup.u.us.sock);
+                       qb_ipcc_us_sock_close(c->setup.u.us.sock);
                        c->setup.u.us.sock = -1;
                }
        }
Comment 6 Ken Gaillot 2015-09-24 15:02:09 UTC 
Patch looks correct. The dispatch_del() function is provided by the client app, and it's a bit of a cruel prank to pass them a closed file descriptor. :)
Comment 7 Christine Caulfield 2015-09-24 15:08:28 UTC 
commit 0f68c8c8283786c271624a3955c8adac878191f9
Author: Christine Caulfield <ccaulfie>
Date:   Thu Sep 24 16:04:28 2015 +0100

    ipc: Don't send the dispatch_del() function a closed fd
Comment 9 Christine Caulfield 2016-01-28 10:38:38 UTC 
This will be covered by the rebase BZ #1299968
Comment 10 Mike McCune 2016-03-28 22:23:55 UTC 
This bug was accidentally moved from POST to MODIFIED via an error in automation, please see mmccune with any questions
Comment 15 errata-xmlrpc 2016-11-04 04:40:27 UTC 
Since the problem described in this bug report should be
resolved in a recent advisory, it has been closed with a
resolution of ERRATA.

For information on the advisory, and where to find the updated
files, follow the link below.

If the solution does not work for you, open a new bug report.

https://rhn.redhat.com/errata/RHBA-2016-2363.html