Use randrw in test_heal

Remove double write when old and new locations are in the same metadata block
Also add another metadata entry fool-safety check which, ideally, will never fire %)
2023-06-03 00:49:53 +03:00 · 2023-06-03 00:47:10 +03:00 · 2023-06-03 00:44:52 +03:00
3 changed files with 22 additions and 13 deletions
--- a/src/blockstore_flush.cpp
+++ b/src/blockstore_flush.cpp
@ -536,14 +536,27 @@ resume_1:
                return false;
            }
            // zero out old metadata entry
+            {
+                clean_disk_entry *old_entry = (clean_disk_entry*)((uint8_t*)meta_old.buf + meta_old.pos*bs->dsk.clean_entry_size);
+                if (old_entry->oid.inode != 0 && old_entry->oid != cur.oid)
+                {
+                    printf("Fatal error (metadata corruption or bug): tried to wipe metadata entry %lu (%lx:%lx v%lu) as old location of %lx:%lx\n",
+                        old_clean_loc >> bs->dsk.block_order, old_entry->oid.inode, old_entry->oid.stripe,
+                        old_entry->version, cur.oid.inode, cur.oid.stripe);
+                    exit(1);
+                }
+            }
            memset((uint8_t*)meta_old.buf + meta_old.pos*bs->dsk.clean_entry_size, 0, bs->dsk.clean_entry_size);
-            await_sqe(15);
-            data->iov = (struct iovec){ meta_old.buf, bs->dsk.meta_block_size };
-            data->callback = simple_callback_w;
-            my_uring_prep_writev(
-                sqe, bs->dsk.meta_fd, &data->iov, 1, bs->dsk.meta_offset + bs->dsk.meta_block_size + meta_old.sector
-            );
-            wait_count++;
+            if (meta_old.sector != meta_new.sector)
+            {
+                await_sqe(15);
+                data->iov = (struct iovec){ meta_old.buf, bs->dsk.meta_block_size };
+                data->callback = simple_callback_w;
+                my_uring_prep_writev(
+                    sqe, bs->dsk.meta_fd, &data->iov, 1, bs->dsk.meta_offset + bs->dsk.meta_block_size + meta_old.sector
+                );
+                wait_count++;
+            }
        }
        if (has_delete)
        {
--- a/src/blockstore_read.cpp
+++ b/src/blockstore_read.cpp
@ -138,10 +138,6 @@ int blockstore_impl_t::dequeue_read(blockstore_op_t *read_op)
        {
            dirty_entry& dirty = dirty_it->second;
            bool version_ok = !IS_IN_FLIGHT(dirty.state) && read_op->version >= dirty_it->first.version;
-            if (IS_SYNCED(dirty.state))
-            {
-                version_ok = true;
-            }
            if (version_ok)
            {
                if (IS_DELETE(dirty.state))
--- a/tests/test_heal.sh
+++ b/tests/test_heal.sh
@ -46,8 +46,8 @@ kill_osds()
 kill_osds &

 LD_PRELOAD="build/src/libfio_vitastor.so" \
-    fio -thread -name=test -ioengine=build/src/libfio_vitastor.so -bs=4k -direct=1 -iodepth=16 -fsync=256 -rw=randwrite \
-        -refill_buffers=1 -mirror_file=./testdata/mirror.bin -etcd=$ETCD_URL -image=testimg -loops=10 -runtime=120
+    fio -thread -name=test -ioengine=build/src/libfio_vitastor.so -bsrange=4k-128k -direct=1 -iodepth=32 -fsync=256 -rw=randrw \
+        -randrepeat=0 -refill_buffers=1 -mirror_file=./testdata/mirror.bin -etcd=$ETCD_URL -image=testimg -loops=10 -runtime=120

 qemu-img convert -S 4096 -p \
    -f raw "vitastor:etcd_host=127.0.0.1\:$ETCD_PORT/v3:image=testimg" \
Author	SHA1	Message	Date
Vitaliy Filippov	9c405009f3	Use randrw in test_heal Test / test_change_pg_count (push) Successful in 33s Details Test / test_change_pg_count_ec (push) Successful in 32s Details Test / test_change_pg_size (push) Successful in 8s Details Test / test_create_nomaxid (push) Successful in 9s Details Test / test_etcd_fail (push) Successful in 51s Details Test / test_failure_domain (push) Successful in 12s Details Test / test_interrupted_rebalance (push) Successful in 1m4s Details Test / test_interrupted_rebalance_imm (push) Successful in 1m0s Details Test / test_interrupted_rebalance_ec (push) Successful in 1m20s Details Test / test_interrupted_rebalance_ec_imm (push) Successful in 55s Details Test / test_minsize_1 (push) Successful in 22s Details Test / test_move_reappear (push) Successful in 22s Details Test / test_rebalance_verify (push) Successful in 1m47s Details Test / test_rebalance_verify_imm (push) Successful in 1m45s Details Test / test_rebalance_verify_ec (push) Successful in 2m3s Details Test / test_rebalance_verify_ec_imm (push) Successful in 1m52s Details Test / test_rm (push) Successful in 11s Details Test / test_snapshot (push) Successful in 15s Details Test / test_snapshot_ec (push) Successful in 16s Details Test / test_splitbrain (push) Successful in 12s Details Test / test_write (push) Successful in 31s Details Test / test_write_xor (push) Successful in 47s Details Test / test_write_no_same (push) Successful in 11s Details Test / test_heal_pg_size_2 (push) Successful in 3m12s Details Test / test_scrub (push) Successful in 24s Details Test / test_scrub_zero_osd_2 (push) Successful in 26s Details Test / test_scrub_xor (push) Successful in 26s Details Test / test_scrub_pg_size_3 (push) Successful in 25s Details Test / test_scrub_pg_size_6_pg_minsize_4_osd_count_6_ec (push) Successful in 42s Details Test / test_scrub_ec (push) Successful in 33s Details	2023-06-03 00:49:53 +03:00
Vitaliy Filippov	f9fbea25a4	Remove double write when old and new locations are in the same metadata block Also add another metadata entry fool-safety check which, ideally, will never fire %)	2023-06-03 00:47:10 +03:00
Vitaliy Filippov	2c9a10d081	Fix an idiotic bug leading to failed reads with -ERANGE with EC :D	2023-06-03 00:44:52 +03:00