Emmanuel Dreyfus <manu@xxxxxxxxxx> wrote: > It does not happen anymore. I think it was a consequence of the other > bug I fixed. I got it again. I have hundreds of warning like this on one brick, then later on the other one: [2012-05-31 21:38:38.952347] W [posix-handle.c:487:posix_handle_hard] 0-pfs-posix: mismatching ino/dev between file /export/wd3a/manu/netbsd/usr/src/tools/binutils/build/opcodes/.libs/libo pcodes.la (58123877/16) and handle /export/wd3a/.glusterfs/a7/e1/a7e1c18f-3695-4f3b-8e2d-948cea0387b1 (58123875/16) Sometimes the warning comes with an error [2012-05-31 21:45:58.268251] E [posix.c:1277:posix_symlink] 0-pfs-posix: setting gfid on /export/wd3a/manu/netbsd/usr/src/tooldir.NetBSD-6.99.4-i386/include/nbin clude/inst.04050a failed[2012-05-31 21:47:05.580996] W [posix-handle.c:461:posix_handle_hard] 0-pfs-posix: link /export/wd3a/manu/netbsd/usr/src/tooldir.NetBSD-6.99.4-i386/include/nbin clude/inst.03423a -> /export/wd3a/.glusterfs/11/c6/11c6f2d4-e4e1-4d02-8fd7-afb5f820fa66 failed (Operation not permitted) I suspect it is memory related: the glusterfsd server-side memory usage grew beyond 900 MB. On the client side, memory got scarce enough that some processes sometime get ENOMEM. -- Emmanuel Dreyfus http://hcpnet.free.fr/pubz manu@xxxxxxxxxx
