Hi Andy,
just to clarify, the thing that Sebastien is referring to is the
maximal buffer size that XrdXrootdXeq and friends send to the OSS AIO
interface. It seems to be set to 2MB max, whereas the optimal results
for the CEPH backend seem to be achieved at around 32MB.
The behavior of the polling threads in the server seems a bit
worrisome too. Netstat shows that there is quite a bit of data stuck
there, so this may be worth tweaking too.
Cheers,
Lukasz
On 03/12/2015 03:43 PM, Sebastien Ponce wrote:
> Hi Andy,
>
> independently of the xattr discussion, I've found another issue that
> prevents me to be able to do fast transfers using the asynchronous
> protocol of xroot.
> The issue (as I understand it) is that the buffer size for each write
> is limited to 2M, due to this line in XrdBuffManager::Recalc :
> // Make sure the request is within our limits
> if (sz <= 0 || sz > maxsz) return 0;
> with this definition of maxsz in the XrdBuffManager constructor :
> maxsz(1<<(XRD_BUSHIFT+XRD_BUCKETS-1+5))
> and these hardcoded values in XrdBuffer.hh :
> #define XRD_BUCKETS 12
> #define XRD_BUSHIFT 10
>
> This gives me a maximum transfer rate of 60 to 70MB/s, limited by the
> ability of the xrootd server to read from the network (the TCP buffers
> are full).
>
> However, recompiling xrootd with XRD_BUCKETS 16 and using 32MB buffers
> gave 300MB/s and XRD_BUCKETS 17 with 64MB buffers gave >500MB/s (limit
> seemed to have been somewhere else then).
>
> So could that value be changed, or at least changeable ?
>
> Cheers,
>
> Sebastien
>
>
########################################################################
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
