On 18/01/16 18:18, Assaf Gordon wrote: > Sorry for the last-minute post, but found 4 failures that might relate to > GPFS file-system ( > https://en.wikipedia.org/wiki/IBM_General_Parallel_File_System ). > > The system is > CentOS 7 > kernel 3.10.0-229.el7.x86_64 > gcc version 4.8.3 20140911 (Red Hat 4.8.3-9) (GCC) > > $ df --output=source,size,used,avail,pcent,target,fstype -h . > Filesystem Size Used Avail Use% Mounted on Type > /dev/commons 2.8P 2.2P 598T 79% /gpfs/commons gpfs > > tested with > coreutils-8.24.163-3f02d > > The failed tests are: > FAIL: tests/tail-2/F-headers > FAIL: tests/tail-2/retry
General problem with tail really on any remote file system. tail: 'a' has been replaced with a remote file. giving up on this name We should probably disable inotify if there are no files opened, as otherwise tail -F is unusable on remote file systems in that use case. I'll do further testing on NFS. > FAIL: tests/misc/shred-passes We ran out of simulated random data (maybe due to extra buffering or something): shred: 'Us': end of file I'll look into increasing it. > FAIL: tests/cp/preserve-slink-time -2016-01-18 12:21:28.294123000 -0500 +2016-01-18 12:21:28.296528000 -0500 cp -Pp didn't seem to preserve the timestamp on GPFS. I'd need access to the system to see why. I'd test with touch -d '...' to see if the timestamp was honored thanks, Pádraig
