From 923e7f5ce5ed437af15e178299a61029ff48e4a2 Mon Sep 17 00:00:00 2001 From: Sage Weil Date: Tue, 31 Jan 2017 15:14:59 -0500 Subject: [PATCH] osd: do not send ENXIO on misdirected op by default In practice this tends to get bubbled up the stack as an error on the caller, and they usually do not handle it properly. For example, with librbd, this turns into EIO and break the VM. Instead, this will manifest as a hung op on the client. That is also not ideal, but given that the root cause here is generally a bug, it's not clear what else would be better. We already log an error in the cluster log, so teuthology runs will continue to fail. Signed-off-by: Sage Weil --- PendingReleaseNotes | 8 ++++++++ src/common/config_opts.h | 1 + src/osd/OSD.cc | 31 ++++++++++++++++++------------- 3 files changed, 27 insertions(+), 13 deletions(-) diff --git a/PendingReleaseNotes b/PendingReleaseNotes index 65347cc2b263e..99a1d9e3b7d0d 100644 --- a/PendingReleaseNotes +++ b/PendingReleaseNotes @@ -30,3 +30,11 @@ either the IoCtx methods on older librados versions or the deprecated methods on any version of librados will lead to incomplete results if/when the new OSD limits are enabled. + +* In previous versions, if a client sent an op to the wrong OSD, the OSD + would reply with ENXIO. The rationale here is that the client or OSD is + clearly buggy and we want to surface the error as clearly as possible. + We now only send the ENXIO reply if the osd_enxio_on_misdirected_op option + is enabled (it's off by default). This means that a VM using librbd that + previously would have gotten an EIO and gone read-only will now see a + blocked/hung IO instead. diff --git a/src/common/config_opts.h b/src/common/config_opts.h index ff8125ef48040..8529fd0c018c7 100644 --- a/src/common/config_opts.h +++ b/src/common/config_opts.h @@ -853,6 +853,7 @@ OPTION(osd_debug_reject_backfill_probability, OPT_DOUBLE, 0) OPTION(osd_debug_inject_copyfrom_error, OPT_BOOL, false) // inject failure during copyfrom completion OPTION(osd_debug_randomize_hobject_sort_order, OPT_BOOL, false) OPTION(osd_debug_misdirected_ops, OPT_BOOL, false) +OPTION(osd_enxio_on_misdirected_op, OPT_BOOL, false) OPTION(osd_debug_verify_cached_snaps, OPT_BOOL, false) OPTION(osd_enable_op_tracker, OPT_BOOL, true) // enable/disable OSD op tracking OPTION(osd_num_op_tracker_shard, OPT_U32, 32) // The number of shards for holding the ops diff --git a/src/osd/OSD.cc b/src/osd/OSD.cc index 1359c8b2956aa..577ed5de5e29a 100644 --- a/src/osd/OSD.cc +++ b/src/osd/OSD.cc @@ -1434,11 +1434,14 @@ void OSDService::handle_misdirected_op(PG *pg, OpRequestRef op) dout(7) << *pg << " misdirected op in " << m->get_map_epoch() << dendl; clog->warn() << m->get_source_inst() << " misdirected " << m->get_reqid() - << " pg " << m->get_pg() - << " to osd." << whoami - << " not " << pg->acting - << " in e" << m->get_map_epoch() << "/" << osdmap->get_epoch() << "\n"; - reply_op_error(op, -ENXIO); + << " pg " << m->get_pg() + << " to osd." << whoami + << " not " << pg->acting + << " in e" << m->get_map_epoch() << "/" << osdmap->get_epoch() + << "\n"; + if (g_conf->osd_enxio_on_misdirected_op) { + reply_op_error(op, -ENXIO); + } } @@ -8785,14 +8788,16 @@ void OSD::handle_op(OpRequestRef& op, OSDMapRef& osdmap) if (!send_map->osd_is_valid_op_target(pgid.pgid, whoami)) { dout(7) << "we are invalid target" << dendl; clog->warn() << m->get_source_inst() << " misdirected " << m->get_reqid() - << " pg " << m->get_pg() - << " to osd." << whoami - << " in e" << osdmap->get_epoch() - << ", client e" << m->get_map_epoch() - << " pg " << pgid - << " features " << m->get_connection()->get_features() - << "\n"; - service.reply_op_error(op, -ENXIO); + << " pg " << m->get_pg() + << " to osd." << whoami + << " in e" << osdmap->get_epoch() + << ", client e" << m->get_map_epoch() + << " pg " << pgid + << " features " << m->get_connection()->get_features() + << "\n"; + if (g_conf->osd_enxio_on_misdirected_op) { + service.reply_op_error(op, -ENXIO); + } return; } -- 2.39.5