oss2.exceptions.PositionNotEqualToLength

[houxiyao]

https://github.com/fsspec/filesystem_spec/blob/73de5045f95257af153769984312c1a4875087b4/fsspec/spec.py#L1491-L1495 https://github.com/fsspec/ossfs/blob/d1db90caee05a359fcd2e59eb3a58f5755495dbf/ossfs/core.py#L746-L748

[houxiyao]

https://help.aliyun.com/document_detail/88432.htm?spm=a2c4g.11186623.0.0.7f602cb7Y726I0#concept-88432-zh

[houxiyao]

Pandas通过fsspec上传大文件是触发了PositionNotEqualToLength

[karajan1001]

Excuse me, could you please provide some more details? how to reproduce this error? Uploading large files, and what does the Pandas do here?

[rgtjf]

I use fsspec with zstd compressor, and encounter the same issues when the object size raises to 5M.

python=3.10 and ossfs=2023.6.0

import json
import fsspec
import ossfs
import sys

#import ossfs.base as ob
#ob.DEFAULT_BLOCK_SIZE = 5 * 2 ** 30
import random
import string

N = 2**20
res = ''.join(random.choices(string.ascii_uppercase + string.digits, k=N))
print(sys.getsizeof(res) / 2**20)

with fsspec.open("oss://personal-cn/junfeng/t1.jsonl.zst", mode="wb", compression="zstd", encoding="utf-8") as f: 
    lines = ''
    for i in range(100):
        line = ''.join(random.choices(string.ascii_uppercase + string.digits, k=N))
        lines += line
        f.write(line.encode("utf-8"))
        print(i, sys.getsizeof(lines) / 2**20)

OSError: [Errno 5] {'status': 409, 'x-oss-request-id': '6541DDD3D832763531ABC4C5', 'details': {'Code': 'PositionNotEqualToLength', 'Message': 'Position is not equal to file length', 'RequestId': '6541DDD3D832763531ABC4C5', 'HostId': 'personal-cn.oss-cn-wulanchabu-internal.aliyuncs.com', 'EC': '0026-00000016', 'RecommendDoc': 'https://api.aliyun.com/troubleshoot?q=0026-00000016'}}

Last, the size of the stored file is about 5.1M.

[rgtjf]

When changing self.loc to self.offset it would be ok.

[idling11]

When changing self.loc to self.offset it would be ok.

it looks like the same error with issue 127