search for: force_readon

After a retry, if the second connection has fewer permissions than the first, but we blindly call next_ops->FOO, we end up triggering an assertion failure in backend.c. This is particularly noticeable when the force_readonly flag is in effect, as that makes it much easier for there to be fewer permissions than before. Add testsuite coverage of pwrite to demonstrate. Signed-off-by: Eric Blake <eblake at redhat.com> --- filters/retry/retry.c | 58 ++++++++++++++++++++++++++++++++++++++----- 1 file changed, 52...

Includes a rework of the previously posted patch for --run improvements (mostly with improved comments and commit message; I decided that waiting for the captive nbdkit to exit was overkill), and four new patches. The tests are intentionally separate, to allow rearranging the order of the series to see the failures being fixed. Eric Blake (6): server: Propagate unexpected nbdkit failure with

...clude <inttypes.h> +#include <string.h> +#include <sys/time.h> + +#include <nbdkit-filter.h> + +#include "cleanup.h" + +static int retries = 5; /* 0 = filter is disabled */ +static int initial_delay = 2; +static bool exponential_backoff = true; +static bool force_readonly = false; + +static int +retry_config (nbdkit_next_config *next, void *nxdata, + const char *key, const char *value) +{ + int r; + + if (strcmp (key, "retries") == 0) { + if (sscanf (value, "%d", &retries) != 1 || retries < 0) { + nbdkit_error (&qu...

...clude <inttypes.h> +#include <string.h> +#include <sys/time.h> + +#include <nbdkit-filter.h> + +#include "cleanup.h" + +static int retries = 5; /* 0 = filter is disabled */ +static int initial_delay = 2; +static bool exponential_backoff = true; +static bool force_readonly = false; + +static int +retry_config (nbdkit_next_config *next, void *nxdata, + const char *key, const char *value) +{ + int r; + + if (strcmp (key, "retries") == 0) { + if (sscanf (value, "%d", &retries) != 1 || retries < 0) { + nbdkit_error (&qu...

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.

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

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.

...t tries to preserve read/write if a read operation fails? > + > +static int retries = 5; /* 0 = filter is disabled */ > +static int initial_delay = 2; Would need a different unit here if you allow sub-second retry. > +static bool exponential_backoff = true; > +static bool force_readonly = false; Initializing a static variable to 0 is sometimes worth avoiding (as it can force .data instead of .bss for a slightly larger binary), but here it makes sense. > + > +static int > +retry_config (nbdkit_next_config *next, void *nxdata, > + const char *key, const...

...= NULL) { + nbdkit_error ("strdup: %m"); + free (h); + return NULL; + } h->reopens = 0; h->open = true; @@ -198,7 +206,8 @@ do_retry (struct retry_handle *h, /* Reopen the connection. */ h->reopens++; - if (next_ops->reopen (nxdata, h->readonly || force_readonly) == -1) { + if (next_ops->reopen (nxdata, + h->readonly || force_readonly, h->exportname) == -1) { /* If the reopen fails we treat it the same way as a command * failing. */ diff --git a/filters/tar/tar.c b/filters/tar/tar.c index a2b26935..c04f09...

...@@ -44,8 +44,8 @@ #include "cleanup.h" -static int retries = 5; /* 0 = filter is disabled */ -static int initial_delay = 2; +static unsigned retries = 5; /* 0 = filter is disabled */ +static unsigned initial_delay = 2; static bool exponential_backoff = true; static bool force_readonly = false; @@ -67,15 +67,15 @@ retry_config (nbdkit_next_config *next, void *nxdata, int r; if (strcmp (key, "retries") == 0) { - if (sscanf (value, "%d", &retries) != 1 || retries < 0) { - nbdkit_error ("cannot parse retries: %s", value); +...

On Mon, Sep 23, 2019 at 12:05:11PM -0500, Eric Blake wrote: > > + int nbdkit_parse_long (const char *what, const char *str, long *r); > > + int nbdkit_parse_unsigned_long (const char *what, > > + const char *str, unsigned long *r); > > Do we really want to encourage the use of parse_long and > parse_unsigned_long? Those differ between

...@@ -44,8 +44,8 @@ #include "cleanup.h" -static int retries = 5; /* 0 = filter is disabled */ -static int initial_delay = 2; +static unsigned retries = 5; /* 0 = filter is disabled */ +static unsigned initial_delay = 2; static bool exponential_backoff = true; static bool force_readonly = false; @@ -67,15 +67,15 @@ retry_config (nbdkit_next_config *next, void *nxdata, int r; if (strcmp (key, "retries") == 0) { - if (sscanf (value, "%d", &retries) != 1 || retries < 0) { - nbdkit_error ("cannot parse retries: %s", value); +...