Print

Print


I'm not convinced this is a bug in gfal.

The xroot-TPC protocol allows the TPC-client (gfal) to tell the destination (dCache) the expected file size.  dCache uses this value to check the data it has received.

The error message from dCache is `File size mismatch (expected=1000000000, actual=1002438656)`.  This means dCache believes the file should be 10^9 bytes long (1 GB), but ended up with more data than this.

I've confirmed that 10^9 is the correct value with a simple `stat`:

```
paul@sprocket:~$ xrdfs ceph-gw8.gridpp.rl.ac.uk:1094 stat /dteam:test1/domatest/jwalder/HTTP_1GB
Path:   /dteam:test1/domatest/jwalder/HTTP_1GB
Id:     2052
Size:   1000000000
MTime:  2021-08-05 12:33:54
Flags:  56 (Offline|IsReadable|IsWritable)
paul@sprocket:~$ 
```
I also downloaded the file with xrdcp, the checked the resulting file.  It is also 10^9 bytes:
```
paul@sprocket:~$ xrdcp root://ceph-gw8.gridpp.rl.ac.uk:1094/dteam:test1/domatest/jwalder/HTTP_1GB .
[953.7MB/953.7MB][100%][==================================================][1.541MB/s]  
paul@sprocket:~$ ls -l HTTP_1GB 
-rw-r--r-- 1 paul paul 1000000000 Aug  5 14:44 HTTP_1GB
paul@sprocket:~$ 
```
I think @abh3 has a point, though.  It looks very much like dCache (for whatever reason) is writing all of the data from the final read, rather than the reduced size.

It is rather odd that this problem hasn't come up before.  Perhaps there is something special about RAL's "test gateway" that triggers a bug in dCache's embedded client.  (But that's just speculation on my part.)

-- 
You are receiving this because you are subscribed to this thread.
Reply to this email directly or view it on GitHub:
https://github.com/xrootd/xrootd/issues/1454#issuecomment-893445009

########################################################################
Use REPLY-ALL to reply to list

To unsubscribe from the XROOTD-DEV list, click the following link:
https://listserv.slac.stanford.edu/cgi-bin/wa?SUBED1=XROOTD-DEV&A=1