tools/xenstore: make the internal memory data base the default

Having a file backed data base has the only advantage of being capable
to dump the contents of it while Xenstore is running, and potentially
using less swap space in case the data base can't be kept in memory.

It has the major disadvantage of a huge performance overhead: switching
to keep the data base in memory only speeds up live update of xenstored
with 120000 nodes from 20 minutes to 11 seconds. A complete tree walk
of this configuration will be reduced from 7 seconds to 280 msecs
(measured by "xenstore-control check").

So make the internal memory data base the default and enhance the
"--internal-db" command line parameter to take an optional parameter
allowing to switch the internal data base back to the file based one.

This is part of XSA-419.

Signed-off-by: Juergen Gross <jgross@suse.com>
Reviewed-by: Julien Grall <jgrall@amazon.com>
(cherry picked from commit d174fefa90487ddd25ebc618028f67b2e8a1f795)

diff --git a/tools/helpers/init-xenstore-domain.c b/tools/helpers/init-xenstore-domain.c
index 32689abd7479f87565d6de15c52f0af92fe3f4c1..d080dae5d3b8b710e225d2b530c3df924a18ffca 100644 (file)
--- a/tools/helpers/init-xenstore-domain.c
+++ b/tools/helpers/init-xenstore-domain.c
@@ -214,9 +214,9 @@ static int build(xc_interface *xch)
     }
 
     if ( param )
-        snprintf(cmdline, 512, "--event %d --internal-db %s", rv, param);
+        snprintf(cmdline, 512, "--event %d %s", rv, param);
     else
-        snprintf(cmdline, 512, "--event %d --internal-db", rv);
+        snprintf(cmdline, 512, "--event %d", rv);
 
     dom->cmdline = xc_dom_strdup(dom, cmdline);
     dom->xenstore_domid = domid;

diff --git a/tools/xenstore/xenstored_core.c b/tools/xenstore/xenstored_core.c
index d5b2e59b0db676a5d609073e7735bf9f31142286..9ddbd934f7944c7516e3fdd9088bb0aeca1f15c6 100644 (file)
--- a/tools/xenstore/xenstored_core.c
+++ b/tools/xenstore/xenstored_core.c
@@ -2230,7 +2230,7 @@ static void accept_connection(int sock)
 }
 #endif
 
-static int tdb_flags;
+static int tdb_flags = TDB_INTERNAL | TDB_NOLOCK;
 
 /* We create initial nodes manually. */
 static void manual_node(const char *name, const char *child)
@@ -2537,7 +2537,8 @@ static void usage(void)
 "                          watch-event: time a watch-event is kept pending\n"
 "  -R, --no-recovery       to request that no recovery should be attempted when\n"
 "                          the store is corrupted (debug only),\n"
-"  -I, --internal-db       store database in memory, not on disk\n"
+"  -I, --internal-db [on|off] store database in memory, not on disk, default is\n"
+"                          memory, with \"--internal-db off\" it is on disk\n"
 "  -K, --keep-orphans      don't delete nodes owned by a domain when the\n"
 "                          domain is deleted (this is a security risk!)\n"
 "  -V, --verbose           to request verbose execution.\n");
@@ -2563,7 +2564,7 @@ static struct option options[] = {
        { "quota-soft", 1, NULL, 'q' },
        { "timeout", 1, NULL, 'w' },
        { "no-recovery", 0, NULL, 'R' },
-       { "internal-db", 0, NULL, 'I' },
+       { "internal-db", 2, NULL, 'I' },
        { "keep-orphans", 0, NULL, 'K' },
        { "verbose", 0, NULL, 'V' },
        { "watch-nb", 1, NULL, 'W' },
@@ -2645,7 +2646,8 @@ int main(int argc, char *argv[])
        orig_argc = argc;
        orig_argv = argv;
 
-       while ((opt = getopt_long(argc, argv, "DE:F:HKNPS:t:A:M:Q:q:T:RVW:w:U",
+       while ((opt = getopt_long(argc, argv,
+                                 "DE:F:HI::KNPS:t:A:M:Q:q:T:RVW:w:U",
                                  options, NULL)) != -1) {
                switch (opt) {
                case 'D':
@@ -2679,7 +2681,8 @@ int main(int argc, char *argv[])
                        tracefile = optarg;
                        break;
                case 'I':
-                       tdb_flags = TDB_INTERNAL|TDB_NOLOCK;
+                       if (optarg && !strcmp(optarg, "off"))
+                               tdb_flags = 0;
                        break;
                case 'K':
                        keep_orphans = true;