You were right.

The limit seems to be somewhere related to fuse.
If I add the strace I see the kernel call to be aligned to 128KB:

...
write(4, "\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0"..., 131072) = 131072
read(3, "\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0"..., 131072) = 131072
write(4, "\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0"..., 131072) = 131072
read(3, "\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0"..., 131072) = 131072
write(4, "\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0"..., 131072) = 131072

I tried several fuse_main options:

xrootdfs ... -o max_read=1048576
cat /proc/mounts
# xrootdfs ....user_id=1000,group_id=1000,allow_other,max_read=1048576...

but doing a cp I see the correct alignment for any size <=128KiB.
As far as I grow bigger than that there is something that is limiting the read request size.

Googling around, it could be related to the fact that setting max_read is not enough, and when creating a fuse session the max_pages should be set. But this is out of my expertise.

—
Reply to this email directly, view it on GitHub, or unsubscribe.
You are receiving this because you are subscribed to this thread.

[ { "@context": "http://schema.org", "@type": "EmailMessage", "potentialAction": { "@type": "ViewAction", "target": "https://github.com/xrootd/xrootd/issues/1694#issuecomment-1130288142", "url": "https://github.com/xrootd/xrootd/issues/1694#issuecomment-1130288142", "name": "View Issue" }, "description": "View this Issue on GitHub", "publisher": { "@type": "Organization", "name": "GitHub", "url": "https://github.com" } } ]

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