quota： Full quota but appending to existing file with echo shows no error

[YunhuiChen]

What happened:

root@chaos-k8s-001:~/chenyunhui/juicefs# ./juicefs quota list redis://127.0.0.1:6379/1
2026/03/11 15:39:54.203675 juicefs[4059727] <INFO>: Meta address: redis://127.0.0.1:6379/1 [NewClient@interface.go:627]
2026/03/11 15:39:54.204183 juicefs[4059727] <WARNING>: AOF is not enabled, you may lose data if Redis is not shutdown properly. [checkRedisInfo@info.go:84]
2026/03/11 15:39:54.204240 juicefs[4059727] <INFO>: Ping redis latency: 20.16µs [checkServerConfig@redis.go:4304]
+----------+---------+---------+------+--------+-------+-------+
|   Path   |   Size  |   Used  | Use% | Inodes | IUsed | IUse% |
+----------+---------+---------+------+--------+-------+-------+
| UID:1000 | 1.0 GiB | 1.0 GiB | 100% |      5 |     1 |   20% |
+----------+---------+---------+------+--------+-------+-------+
root@chaos-k8s-001:~/chenyunhui/juicefs# su - user1

using echo to append to an existing file does not report an error:
$ echo 1 >> /juicefs1/test/tt

new file:
$ echo 1 >> /juicefs1/test/tt2
-sh: 2: cannot create /juicefs1/test/tt2: Disk quota exceeded

What you expected to happen:

How to reproduce it (as minimally and precisely as possible):

Anything else we need to know?

Environment:

• JuiceFS version (use juicefs --version) or Hadoop Java SDK version:
• Cloud provider or hardware configuration running JuiceFS:
• OS (e.g cat /etc/os-release):
• Kernel (e.g. uname -a):
• Object storage (cloud provider and region, or self maintained):
• Metadata engine info (version, cloud provider managed or self maintained):
• Network connectivity (JuiceFS to metadata engine, JuiceFS to object storage):
• Others:

[Xuyuchao-juice]

因为现在的write流程是先返回，然后启动goroutine去异步提交写入，判断quota也是在异步任务里面做的，所以表面上命令会返回成功，实际上内容是写不进去的

[YunhuiChen]

在之前的--path相关的CI测试里echo 会直接报错：
case:

test_remove_and_restore(){
    prepare_test
    ./juicefs format $META_URL myjfs
    ./juicefs mount -d $META_URL /jfs --heartbeat $HEARTBEAT_INTERVAL
    mkdir -p /jfs/d
    ./juicefs quota set $META_URL --path /d --capacity 1
    sleep $((HEARTBEAT_INTERVAL+HEARTBEAT_SLEEP))
    dd if=/dev/zero of=/jfs/d/test1 bs=1G count=1
    sleep $DIR_QUOTA_FLUSH_INTERVAL
    ./juicefs quota get $META_URL --path /d 2>&1 | tee quota.log
    used=$(cat quota.log | grep "/d" | awk -F'|' '{print $5}'  | tr -d '[:space:]')
    [[ $used != "100%" ]] && echo "used should be 100%" && exit 1 || true
    echo a | tee -a /jfs/d/test1 2>error.log && echo "write should fail on out of space" && exit 1 || true
    grep -i "Disk quota exceeded" error.log || (echo "grep failed" && exit 1)

1+0 records in
1+0 records out
1073741824 bytes (1.1 GB, 1.0 GiB) copied, 1.06109 s, 1.0 GB/s
warning: GOCOVERDIR not set, no coverage data emitted
2026/03/17 20:57:23.978241 juicefs[11212] <INFO>: Meta address: ***(127.0.0.1)/test?max_open_conns=30 [NewClient@interface.go:627]
+------+---------+---------+------+-----------+-------+-------+
| Path |   Size  |   Used  | Use% |   Inodes  | IUsed | IUse% |
+------+---------+---------+------+-----------+-------+-------+
| /d   | 1.0 GiB | 1.0 GiB | 100% | unchanged |     1 |       |
+------+---------+---------+------+-----------+-------+-------+
a
tee: /jfs/d/test1: Disk quota exceeded

[Xuyuchao-juice]

之前的测试脚本用的写入方式是 tee命令来写入，现在的测试方法用的是echo >>的方式写入。
测了一下确实tee会报错，echo >> 不会。问了下ai的说法是echo只检查write的返回值，tee命令会监控整个（open->write->close）全周期的返回值。

我写了个C程序测了下目录quota，write确实不会报错（因为是异步任务先返回了），close时才报错