Bug#445156: nfsd: non-standard errno: -14 when receiving large amounts of data via nfs
Package: linux-image-2.6.18-5-iop32x
Version: 2.6.18.dfsg.1-13etch1
Severity: important
When transfering large amounts of data from another computer (nfs client) to my
Thecus N2100 (nfs server), I get "Input/output error" on the client and
"nfsd: non-standard errno: -14" in dmesg on the server (the Thecus) and the
target file gets corrupted. The error occurs maybe once per 5 GB of data.
This is the same as
http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=404447
but on a Thecus N2100 (iop32x) instead of a NSLU (ixp4xx). #404447 was
fixed in 2.6.18.dfsg.1-9.
Some more random information:
Transfers in the other direction work without problems
ifconfig shows errors:0 (and it shouldn't matter, since I use tcp which should
correct transfer errors).
The nfs share is on a LVM volume.
# lsmod
Module Size Used by
nfs 128208 0
nfsd 113040 13
exportfs 5792 1 nfsd
lockd 67252 3 nfs,nfsd
nfs_acl 4032 2 nfs,nfsd
ipv6 269640 19
sunrpc 165912 19 nfs,nfsd,lockd,nfs_acl
loop 16200 0
ext3 137704 3
jbd 59368 1 ext3
mbcache 9188 1 ext3
dm_mirror 20948 0
dm_snapshot 17180 0
dm_mod 59120 7 dm_mirror,dm_snapshot
raid1 23488 2
md_mod 80628 3 raid1
ide_generic 1216 0 [permanent]
ide_core 114280 1 ide_generic
sd_mod 19248 8
ehci_hcd 31240 0
sata_sil 12040 6
uhci_hcd 22536 0
usbcore 126600 3 ehci_hcd,uhci_hcd
libata 105460 1 sata_sil
scsi_mod 102800 2 sd_mod,libata
r8169 30732 0
Reply to: