On 1/31/20 9:10 PM, Jay Fenlason wrote:
I have a small FreeIPA deployment of ~6-8 servers running on Centos
7.7. Do to the addition and removal of some of the servers, some
cruft (tombstones, replication conflicts, etc) have crept in to the
directory. I noticed that when I attempted to delete some of the
cruft entries, ns-slapd would hang, failirg to process requests, or
even shut down. In my investigation I found that pre-write callback
of the plugins (including the nis plugin) were being called, but no
corresponding post-write call was happening. The nis plugin takes a
write lock in the pre-write callback and releases it in the post-write
callback. Never releasing the lock causes the next attempt to take it
to hang. I wrote the attached patch, which simply ensures the
post-write callback is always called if the pre-write callback was
called.
I need to look at this closer, but I would suspect you should not ignore
the check for delete_tombstone_entry:
+ if (pre_delete_called) {
To
+ if (pre_delete_called || !delete_tombstone_entry) {
Does it still resolve your problem? Probably, but best to check...
Like I said this needs a closer investigation/testing to make sure it does not
cause any regressions. Do you have a stack trace from the hang you
encountered? If not, could you get one for us?
Thanks,
Mark
This is such a simple patch that it scarcely needs a signoff.
Nonetheless:
Signed-off-by: Jay Fenlason <[email protected]>
--- a/ldap/servers/slapd/back-ldbm/ldbm_delete.c.orig 2020-01-31
07:28:04.085861174 -0500
+++ b/ldap/servers/slapd/back-ldbm/ldbm_delete.c 2020-01-31
07:30:33.932947489 -0500
@@ -81,6 +81,7 @@
Connection *pb_conn;
int32_t parent_op = 0;
struct timespec parent_time;
+ int pre_delete_called = 0;
if (slapi_pblock_get(pb, SLAPI_CONN_ID, &conn_id) < 0) {
conn_id = 0; /* connection is NULL */
@@ -371,6 +372,7 @@
}
if (retval == 0) {
retval = plugin_call_plugins(pb,
SLAPI_PLUGIN_BE_PRE_DELETE_FN);
+ pre_delete_called = 1;
}
if (retval)
{
@@ -1491,7 +1493,7 @@
* The bepostop is called even if the operation fails,
* but not if the operation is purging tombstones.
*/
- if (!delete_tombstone_entry) {
+ if (pre_delete_called) {
plugin_call_plugins(pb, SLAPI_PLUGIN_BE_POST_DELETE_FN);
}
/* Need to return to cache after post op plugins are called */
_______________________________________________
389-devel mailing list -- [email protected]
To unsubscribe send an email to [email protected]
Fedora Code of Conduct:
https://docs.fedoraproject.org/en-US/project/code-of-conduct/
List Guidelines: https://fedoraproject.org/wiki/Mailing_list_guidelines
List Archives:
https://lists.fedoraproject.org/archives/list/[email protected]
--
389 Directory Server Development Team
_______________________________________________
389-devel mailing list -- [email protected]
To unsubscribe send an email to [email protected]
Fedora Code of Conduct:
https://docs.fedoraproject.org/en-US/project/code-of-conduct/
List Guidelines: https://fedoraproject.org/wiki/Mailing_list_guidelines
List Archives:
https://lists.fedoraproject.org/archives/list/[email protected]
