aboutsummaryrefslogtreecommitdiffstats
diff options
context:
space:
mode:
authorBrian Behlendorf <[email protected]>2022-03-08 09:16:35 -0800
committerGitHub <[email protected]>2022-03-08 09:16:35 -0800
commit6df43169b30985ce7d9df11d25093dec26829d2c (patch)
tree54b6fa3bc3bea7b231ba32b8d02d53089d917f74
parent39a4daf742a4bd34b34f85e004895385e4c46c1a (diff)
Fix ENOSPC when unlinking multiple files from full pool
When unlinking multiple files from a pool at 100% capacity, it was possible for ENOSPC to be returned after the first unlink. e.g. rm -f /mnt/fs/test1.0.0 /mnt/fs/test1.1.0 /mnt/fs/test1.2.0 rm: cannot remove '/mnt/fs/test1.1.0': No space left on device rm: cannot remove '/mnt/fs/test1.2.0': No space left on device After waiting for the pending deferred frees from the first unlink to be processed the remaining files can then be unlinked. This is caused by the quota limit in dsl_dir_tempreserve_impl() being temporarily decreased to the allocatable pool capacity less any deferred free space. This is resolved using the existing mechanism of returning ERESTART when over quota as long as we know enough space will shortly be available after processing the pending deferred frees. Reviewed-by: Alexander Motin <[email protected]> Reviewed-by: Ryan Moeller <[email protected]> Reviewed-by: Tony Hutter <[email protected]> Signed-off-by: Brian Behlendorf <[email protected]> Closes #13172
-rw-r--r--include/sys/dsl_pool.h1
-rw-r--r--module/zfs/dsl_dir.c20
-rw-r--r--module/zfs/dsl_pool.c6
-rw-r--r--tests/runfiles/common.run2
-rw-r--r--tests/zfs-tests/tests/functional/no_space/Makefile.am3
-rwxr-xr-xtests/zfs-tests/tests/functional/no_space/enospc_rm.ksh60
6 files changed, 84 insertions, 8 deletions
diff --git a/include/sys/dsl_pool.h b/include/sys/dsl_pool.h
index 0283a8c58..32d973f09 100644
--- a/include/sys/dsl_pool.h
+++ b/include/sys/dsl_pool.h
@@ -162,6 +162,7 @@ int dsl_pool_sync_context(dsl_pool_t *dp);
uint64_t dsl_pool_adjustedsize(dsl_pool_t *dp, zfs_space_check_t slop_policy);
uint64_t dsl_pool_unreserved_space(dsl_pool_t *dp,
zfs_space_check_t slop_policy);
+uint64_t dsl_pool_deferred_space(dsl_pool_t *dp);
void dsl_pool_wrlog_count(dsl_pool_t *dp, int64_t size, uint64_t txg);
boolean_t dsl_pool_wrlog_over_max(dsl_pool_t *dp);
void dsl_pool_dirty_space(dsl_pool_t *dp, int64_t space, dmu_tx_t *tx);
diff --git a/module/zfs/dsl_dir.c b/module/zfs/dsl_dir.c
index 43ef18978..aca32ff9b 100644
--- a/module/zfs/dsl_dir.c
+++ b/module/zfs/dsl_dir.c
@@ -1334,13 +1334,21 @@ top_of_function:
/*
* If they are requesting more space, and our current estimate
* is over quota, they get to try again unless the actual
- * on-disk is over quota and there are no pending changes (which
- * may free up space for us).
+ * on-disk is over quota and there are no pending changes
+ * or deferred frees (which may free up space for us).
*/
if (used_on_disk + est_inflight >= quota) {
- if (est_inflight > 0 || used_on_disk < quota ||
- (retval == ENOSPC && used_on_disk < quota))
- retval = ERESTART;
+ if (est_inflight > 0 || used_on_disk < quota) {
+ retval = SET_ERROR(ERESTART);
+ } else {
+ ASSERT3U(used_on_disk, >=, quota);
+
+ if (retval == ENOSPC && (used_on_disk - quota) <
+ dsl_pool_deferred_space(dd->dd_pool)) {
+ retval = SET_ERROR(ERESTART);
+ }
+ }
+
dprintf_dd(dd, "failing: used=%lluK inflight = %lluK "
"quota=%lluK tr=%lluK err=%d\n",
(u_longlong_t)used_on_disk>>10,
@@ -1348,7 +1356,7 @@ top_of_function:
(u_longlong_t)quota>>10, (u_longlong_t)asize>>10, retval);
mutex_exit(&dd->dd_lock);
DMU_TX_STAT_BUMP(dmu_tx_quota);
- return (SET_ERROR(retval));
+ return (retval);
}
/* We need to up our estimated delta before dropping dd_lock */
diff --git a/module/zfs/dsl_pool.c b/module/zfs/dsl_pool.c
index c120d19b8..c87660905 100644
--- a/module/zfs/dsl_pool.c
+++ b/module/zfs/dsl_pool.c
@@ -950,6 +950,12 @@ dsl_pool_unreserved_space(dsl_pool_t *dp, zfs_space_check_t slop_policy)
return (quota);
}
+uint64_t
+dsl_pool_deferred_space(dsl_pool_t *dp)
+{
+ return (metaslab_class_get_deferred(spa_normal_class(dp->dp_spa)));
+}
+
boolean_t
dsl_pool_need_dirty_delay(dsl_pool_t *dp)
{
diff --git a/tests/runfiles/common.run b/tests/runfiles/common.run
index eb0d45135..fdc2172b6 100644
--- a/tests/runfiles/common.run
+++ b/tests/runfiles/common.run
@@ -694,7 +694,7 @@ tags = ['functional', 'nestedfs']
[tests/functional/no_space]
tests = ['enospc_001_pos', 'enospc_002_pos', 'enospc_003_pos',
- 'enospc_df']
+ 'enospc_df', 'enospc_rm']
tags = ['functional', 'no_space']
[tests/functional/nopwrite]
diff --git a/tests/zfs-tests/tests/functional/no_space/Makefile.am b/tests/zfs-tests/tests/functional/no_space/Makefile.am
index c2e42bc2a..31584fb17 100644
--- a/tests/zfs-tests/tests/functional/no_space/Makefile.am
+++ b/tests/zfs-tests/tests/functional/no_space/Makefile.am
@@ -5,7 +5,8 @@ dist_pkgdata_SCRIPTS = \
enospc_001_pos.ksh \
enospc_002_pos.ksh \
enospc_003_pos.ksh \
- enospc_df.ksh
+ enospc_df.ksh \
+ enospc_rm.ksh
dist_pkgdata_DATA = \
enospc.cfg
diff --git a/tests/zfs-tests/tests/functional/no_space/enospc_rm.ksh b/tests/zfs-tests/tests/functional/no_space/enospc_rm.ksh
new file mode 100755
index 000000000..065abc759
--- /dev/null
+++ b/tests/zfs-tests/tests/functional/no_space/enospc_rm.ksh
@@ -0,0 +1,60 @@
+#!/bin/ksh -p
+#
+# CDDL HEADER START
+#
+# This file and its contents are supplied under the terms of the
+# Common Development and Distribution License ("CDDL"), version 1.0.
+# You may only use this file in accordance with the terms of version
+# 1.0 of the CDDL.
+#
+# A full copy of the text of the CDDL should have accompanied this
+# source. A copy of the CDDL is also available via the Internet at
+# http://www.illumos.org/license/CDDL.
+#
+# CDDL HEADER END
+#
+
+#
+# Copyright (c) 2014, 2016 by Delphix. All rights reserved.
+# Copyright (c) 2022 by Lawrence Livermore National Security, LLC.
+#
+
+. $STF_SUITE/include/libtest.shlib
+. $STF_SUITE/tests/functional/no_space/enospc.cfg
+
+#
+# DESCRIPTION:
+# After filling a filesystem, verify the contents can be removed
+# without encountering an ENOSPC error.
+#
+
+verify_runnable "both"
+
+function cleanup
+{
+ destroy_pool $TESTPOOL
+ log_must rm -f $all_vdevs
+}
+
+log_onexit cleanup
+
+log_assert "Files can be removed from full file system."
+
+all_vdevs=$(echo $TEST_BASE_DIR/file.{01..12})
+
+log_must truncate -s $MINVDEVSIZE $all_vdevs
+
+log_must zpool create -f $TESTPOOL draid2:8d:2s $all_vdevs
+log_must zfs create $TESTPOOL/$TESTFS
+log_must zfs set mountpoint=$TESTDIR $TESTPOOL/$TESTFS
+log_must zfs set compression=off $TESTPOOL/$TESTFS
+
+log_note "Writing files until ENOSPC."
+log_mustnot_expect "No space left on device" fio --name=test \
+ --fallocate=none --rw=write --bs=1M --size=1G --numjobs=4 \
+ --sync=1 --directory=$TESTDIR/ --group_reporting
+
+log_must rm $TESTDIR/test.*
+log_must test -z "$(ls -A $TESTDIR)"
+
+log_pass "All files removed without error"