journald: don't flush to /var/log/journal before we get asked to

This changes journald to not write to /var/log/journal until it received
SIGUSR1 for the first time, thus having been requested to flush the runtime
journal to disk.

This makes the journal work nicer with systems which have the root file system
writable early, but still need to rearrange /var before journald should start
writing and creating files to it, for example because ACLs need to be applied
first, or because /var is to be mounted from another file system, NFS or tmpfs
(as is the case for systemd.volatile=state).

Before this change we required setupts with /var split out to mount the root
disk read-only early on, and ship an /etc/fstab that remounted it writable only
after having placed /var at the right place. But even that was racy for various
preparations as journald might end up accessing the file system before it was
entirely set up, as soon as it was writable.

With this change we make scheduling when to start writing to /var/log/journal
explicit. This means persistent mode now requires
systemd-journal-flush.service in the mix to work, as otherwise journald would
never write to the directory.

See: #1397

1 files changed, 11 insertions, 10 deletions

diff --git a/src/journal/journald-server.c b/src/journal/journald-server.c
index 6bdb375fb7..8b92ea3def 100644
--- a/src/journal/journald-server.c
+++ b/src/journal/journald-server.c
@@ -283,17 +283,16 @@ static int open_journal(
 }
 
 static bool flushed_flag_is_set(void) {
-        return (access("/run/systemd/journal/flushed", F_OK) >= 0);
+        return access("/run/systemd/journal/flushed", F_OK) >= 0;
 }
 
 static int system_journal_open(Server *s, bool flush_requested) {
-        bool flushed = false;
         const char *fn;
         int r = 0;
 
         if (!s->system_journal &&
-            (s->storage == STORAGE_PERSISTENT || s->storage == STORAGE_AUTO) &&
-            (flush_requested || (flushed = flushed_flag_is_set()))) {
+            IN_SET(s->storage, STORAGE_PERSISTENT, STORAGE_AUTO) &&
+            (flush_requested || flushed_flag_is_set())) {
 
                 /* If in auto mode: first try to create the machine
                  * path, but not the prefix.
@@ -326,8 +325,8 @@ static int system_journal_open(Server *s, bool flush_requested) {
                  * Perform an implicit flush to var, leaving the runtime
                  * journal closed, now that the system journal is back.
                  */
-                if (s->runtime_journal && flushed)
-                        (void) server_flush_to_var(s);
+                if (!flush_requested)
+                        (void) server_flush_to_var(s, true);
         }
 
         if (!s->runtime_journal &&
@@ -1183,7 +1182,7 @@ finish:
         dispatch_message_real(s, iovec, n, m, ucred, tv, label, label_len, unit_id, priority, object_pid);
 }
 
-int server_flush_to_var(Server *s) {
+int server_flush_to_var(Server *s, bool require_flag_file) {
         sd_id128_t machine;
         sd_journal *j = NULL;
         char ts[FORMAT_TIMESPAN_MAX];
@@ -1193,13 +1192,15 @@ int server_flush_to_var(Server *s) {
 
         assert(s);
 
-        if (s->storage != STORAGE_AUTO &&
-            s->storage != STORAGE_PERSISTENT)
+        if (!IN_SET(s->storage, STORAGE_AUTO, STORAGE_PERSISTENT))
                 return 0;
 
         if (!s->runtime_journal)
                 return 0;
 
+        if (require_flag_file && !flushed_flag_is_set())
+                return 0;
+
         (void) system_journal_open(s, true);
 
         if (!s->system_journal)
@@ -1411,7 +1412,7 @@ static int dispatch_sigusr1(sd_event_source *es, const struct signalfd_siginfo *
 
         log_info("Received request to flush runtime journal from PID " PID_FMT, si->ssi_pid);
 
-        (void) server_flush_to_var(s);
+        (void) server_flush_to_var(s, false);
         server_sync(s);
         server_vacuum(s, false);