Cannot write files to filesystem, commit 933ec99 (commit 986dd8a works) #5720
Comments
|
Please post the config.log for zfs. |
|
OK I found the bug. It would affect kernel version 3.16-4.0 |
|
Try this #5726 |
|
The Debian builder should have caught this, it's running a 3.16.0-4-amd64 kernel. Also related to these changes see comment 933ec99#commitcomment-20703820 |
|
We're only doing build tests on Debian we don't run the full test suite. There must not be any other kernels in that range running the full suite. |
The .write/.read file operations callbacks can be retired since support for .read_iter/.write_iter and .aio_read/.aio_write has been added. The vfs_write()/vfs_read() entry functions will select the correct interface for the kernel. This is desirable because all VFS write/read operations now rely on common code. This change also add the generic write checks to make sure that ulimits are enforced correctly on write. Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov> Signed-off-by: Chunwei Chen <david.chen@osnexus.com> Closes #5587 Closes #5673
|
@tuxoko As requested, here the config on my fedora backup box 4.9.6-200.fc25.x86_64 now running 986dd8a ("OpenZFS 5561 - support root pools on EFI/GPT partitioned disks", 2017-01-27) There is not much done on that box, it runs 2x2TB HDDs, and an old SSD for a xfs-over-bcache-over-zvol (thanks to yet non-persistent l2arc). So in a way it is good that the federa zfs packages and/or my dkms failed working, I come to contribute to zfs ;-) Although I had a different mind tonight... on my manjaro the tortoishg is broke thanks to qscintilla bindings for python not updated. So the missing zpool (again) on the backup box is rather 2nd prio,.. |
|
@grizzlyfred |
|
I thought it was that. I built the 986dd8a ("OpenZFS 5561 - support root pools on EFI/GPT partitioned disks", 2017-01-27) into kmod rpms and installed them, then I git pulled and then I used gitk to hardreset |
|
@grizzlyfred you may want to look at #4365. (Especially if you have some test box. I have a test backup machine up running with a version of it since a month.) It ought to be more potent than persistent l2arc. |
|
@grizzlyfred |
|
@grizzlyfred |
|
Now that you say.. I overlooked that on https://github.com/zfsonlinux/zfs/wiki/Custom-Packages
I never read that text on top, just the commands down. My bad. Now I feel stupid. Yeah autogen.... How could I forget... Will see if that changes anything... |
|
Yeah, you need to rerun ./autogen.sh when you switch version, otherwise new conftest won't get run. |
|
Yep. Current head 0f676dc successfully built. |
For generic_write_checks with 2 args, we can exit when it returns zero because it means count is zero. However this is not the case for generic_write_checks with 4 args, where zero means no error. Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov> Reviewed-by: Haakan T Johansson <f96hajo@chalmers.se> Signed-off-by: Chunwei Chen <david.chen@osnexus.com> Closes openzfs#5720 Closes openzfs#5726
For generic_write_checks with 2 args, we can exit when it returns zero because it means count is zero. However this is not the case for generic_write_checks with 4 args, where zero means no error. Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov> Reviewed-by: Haakan T Johansson <f96hajo@chalmers.se> Signed-off-by: Chunwei Chen <david.chen@osnexus.com> Closes openzfs#5720 Closes openzfs#5726 Requires-builders: style
For generic_write_checks with 2 args, we can exit when it returns zero because it means count is zero. However this is not the case for generic_write_checks with 4 args, where zero means no error. Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov> Reviewed-by: Haakan T Johansson <f96hajo@chalmers.se> Signed-off-by: Chunwei Chen <david.chen@osnexus.com> Closes #5720 Closes #5726
For generic_write_checks with 2 args, we can exit when it returns zero because it means count is zero. However this is not the case for generic_write_checks with 4 args, where zero means no error. Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov> Reviewed-by: Haakan T Johansson <f96hajo@chalmers.se> Signed-off-by: Chunwei Chen <david.chen@osnexus.com> Closes openzfs#5720 Closes openzfs#5726
For generic_write_checks with 2 args, we can exit when it returns zero because it means count is zero. However this is not the case for generic_write_checks with 4 args, where zero means no error. Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov> Reviewed-by: Haakan T Johansson <f96hajo@chalmers.se> Signed-off-by: Chunwei Chen <david.chen@osnexus.com> Closes openzfs#5720 Closes openzfs#5726
System information
Describe the problem you're observing
Writes fail with commit 933ec99, reporting that the filesystem is full. With commit 986dd8a it works.
Describe how to reproduce the problem
the last command reports failure:
Commit 933ec99 is for PR #5673 . Is the problem some kind of kernel configuration with my (old?) kernel?
I did redo
after checkout of 933ec99. (And same with 986dd8a that works.)
The text was updated successfully, but these errors were encountered: