similar to: [nbdkit PATCH v2 0/6] Improve retry filter

v1 was here: https://www.redhat.com/archives/libguestfs/2019-September/msg00199.html v2: - Adds a fairly simple yet comprehensive test using sh plugin. - Rebase and retest. Patch 1 is a misc patch not really related to the series. Rich.

This is a retry filter implementation as outlined here: https://www.redhat.com/archives/libguestfs/2019-September/msg00167.html It is only lightly tested. One way to test it is to try an SSH copy (see the commit message for patch 2/2), and in the middle of the copy kill the per-connection sshd on the remote machine. You will see that the copy recovers after a few seconds. Add the nbdkit -v

v2 was here: https://www.redhat.com/archives/libguestfs/2019-September/msg00221.html I think this is more like "the one". It handles reopen failing correctly, and there is a second test for that. I also ran my sshd tests locally and it worked in all scenarios I could think up (except of course sshd not being available at the start, but we want that to fail). Rich.

I think this is my last round of patches for issues I identified with the retry filter. With this in place, it should be safe to interject another filter in between retry and the plugin. Eric Blake (5): retry: Don't call into closed plugin tests: Refactor test-retry-reopen-fail.sh tests: Enhance retry test to cover failed reopen server: Move prepare/finalize/close recursion to

For review only, this needs some clean up and more tests. My eyes are going cross-eyed looking at the calculate_offset_of_entry function, so time to take a break ... Rich.

I'm still working on another set of patches to have reopen call .finalize/.prepare (so that another filter can safely appear between retry and the plugin), but for tonight, these are the patches I think are ready to go. Eric Blake (4): retry: Handle can_fua and can_fast_zero changes tests: Test retry with different fua/fast-zero flags server: Close backends if a filter's .open fails

There's more that can be done (in particular, use of CLEANUP_FREE), but this is enough to at least see if I'm on the right track. I couldn't figure out an obvious difference between common/include and common/utils, but it looks like the former is for things that are inlineable via .h only, while the latter is when you need to link in a convenience library, so this landed in the

I still don't have .prepare/.finalize working cleanly across reopen, but did find a nasty bug where a botched assertion means we failed to notice reads beyond EOF in both the xz and retry filter. Refactoring backend.c will make .finalize work easier. Eric Blake (5): xz: Avoid reading beyond EOF retry: Check size before transactions tests: Test retry when get_size values change

In https://bugzilla.redhat.com/show_bug.cgi?id=1841820, it was pointed out that the retry filter not retrying .open means that an ssh connection (such as in a vmx+ssh v2v conversion) fails when the ssh connection itself cannot be retried. A year ago, this was an inherent limitation of our retry implementation; but in the meantime, my work to allow filters to open independent backends has made it

Ultimately, both the blocksize and swab filters want to return aligned extents to the client. I'm posting this as a snapshot of my work in progress on how I plan to get there (it's not quite working yet, but I'm done for today and wanted to at least document my ideas). I might also add a convenience function for nbdkit_extents_offset, since we have a number of filters that repeat the

When querying the extents of the underlying plugin, we should only translate the starting offset, and let the plugin report for at least as many bytes as our range permits. Otherwise, short-changing the range causes bad behavior such as returning 0 extents, or even failing the creation of an extents tracker: $ cat script case "$1" in get_size) echo 1m;; can_extents) ;;

This has already been pushed upstream. I am simply posting these here so we have a reference in the mailing list in case we find bugs later (as I'm sure we will - it's a complex patch series). Great thanks to Eric Blake for tireless review on this one. It also seems to have identified a few minor bugs in qemu along the way. Rich.

This fix isn't exhaustive but it fixes some obvious problems in the filters. Rich.

I'm not sure exactly which version we're up to, but let's say it's version 4. I'm a lot happier with this version: - all filters have been reviewed and changed where I think that's necessary - can_extents is properly defined and implemented now - NBD protocol is followed - I believe it addresses all previous review points where possible The "only" thing

On Thu, Oct 03, 2019 at 09:54:37PM -0500, Eric Blake wrote: > Although it is unusual for a plugin to shrink size on reopen, it is > not impossible. Failure to check our bounds could result in violating > assumptions in the plugin that all requests are in-bounds. Note that > if the plugin gains a larger size, we merely never access the new tail > of the file (when the NBD protocol

Not sure what version we're up to, but this reimplements extents using the new simpler structure described in this thread: https://www.redhat.com/archives/libguestfs/2019-March/msg00077.html I also fixed most of the things that Eric pointed out in the previous review, although I need to go back over his replies and check I've got everything. This needs a bit more testing. However the

While working on adding assertions to pthread_mutex_lock calls, I noticed that the truncate filter's use of mutex didn't really protect us, and isn't really necessary. Cleaning that up also spotted a couple of other potential cleanups. Eric Blake (4): filters: Drop useless .open callbacks truncate: Fix corruption when plugin changes per-connection size truncate: Test for safe

Now that a filter can open a backend as many times as it wants, there's no longer a technical reason we can't retry .open. However, adding retry logic here does mean we have to weaken an assert in the server backend code, since prepare can now be reached more than once. Test coverage will be added in a separate patch, so that it becomes easy to swap patch order and see that the test

On 4/23/19 2:06 PM, Eric Blake wrote: > Now that cleanup.h is in common code, we can use it in our > filters. The first round focuses just on places that called > nbdkit_extents_free(), as all three callers had multiple exit paths > that definitely benefit from the macro. > > Signed-off-by: Eric Blake <eblake@redhat.com> > --- > filters/offset/offset.c | 13

I do have a question about whether patch 2 is right, or whether I've exposed a bigger problem in the truncate (and possibly other) filter, but the rest seem fairly straightforward. Eric Blake (4): server: Check for pthread lock failures truncate: Factor out reading real_size under mutex plugins: Check for mutex failures filters: Check for mutex failures filters/cache/cache.c