bug#45887:
Thanks a lot for reply. I think it make sense calling fsync when parted is done with device. but I'm confusing because buffer cache was flushed with 'print' command in parted also. I think 'parted [device partitoin] -s print' is just for getting information about device, not for making any changes to device. Is this also normal?
bug#45887: buffer cache flushed after excute 'parted -s print' command
On Fri, Jan 15, 2021 at 12:32:42PM +0900, Hyun Ha wrote: > Hi, > I have a strange issue with parted. > After execute parted command, buffer cache flushed and I don't know why. > I have nvme device to set bcache and HDD device for backing block device of > bcache and writeback mode. Yes, parted calls fsync when it is done with the device. It always opens the device in RW mode and always fsync's it when closing it to make sure any changes make it to disk. Brian -- Brian C. Lane (PST8PDT) - weldr.io - lorax - parted - pykickstart
bug#45887: buffer cache flushed after excute 'parted -s print' command
Hi, I have a strange issue with parted. After execute parted command, buffer cache flushed and I don't know why. I have nvme device to set bcache and HDD device for backing block device of bcache and writeback mode. System informations: - OS: CentOS Linux release 7.4.1708 (Core) - Kernel version: 4.18.16-1.el7.elrepo.x86_64 - parted version: parted (GNU parted) 3.1 Reproduce: - create nvme partition - make bcache - execute parted command Test: $ free ; parted /dev/nvme0n1p10 -s print;free totalusedfree shared buff/cache available Mem: 1314637882727630874617372306429570108 102167828 Swap: 4194300 0 4194300 Error: /dev/nvme0n1p10: unrecognised disk label Model: Unknown (unknown) Disk /dev/nvme0n1p10: 50.0GB Sector size (logical/physical): 512B/512B Partition Table: unknown Disk Flags: totalusedfree shared buff/cache available Mem: 1314637882727678475484216306428702788 102167836 Swap: 4194300 0 4194300 $ free ; parted /dev/nvme0n1p11 -s print;free totalusedfree shared buff/cache available Mem: 1314637882732402875436544312828703216 102119460 Swap: 4194300 0 4194300 Error: /dev/nvme0n1p11: unrecognised disk label Model: Unknown (unknown) Disk /dev/nvme0n1p11: 50.0GB Sector size (logical/physical): 512B/512B Partition Table: unknown Disk Flags: totalusedfree shared buff/cache available Mem: 1314637882732196876610612312827531208 102121764 Swap: 4194300 0 4194300 gdisk and fdisk does not flush buffer cache with execute command. Is this a bug or normal? thank you in advance.