Redis Redis-cli 6.0.1 --cluster-yes doesn't work

Hi, I have been moving a Redis Cluster to Redis 6.0.1 and it seems that the --cluster-yes option is not working for redis-cli 6.0.1 as before. I version 5, setting --cluster-yes allowed me to skip the question to proceed to create the cluster. But now it asks to proceed even using that option. Just to clarify I am talking about this question:

...
Can I set the above configuration? (type 'yes' to accept):

As a workaround we did:

yes yes | redis-cli --cluster create ...

And it solves the issue temporarily, but I wanted to share/report this bug.

Comment From: bsergean

The code that is about to start the node creation is:

    if (confirmWithYes("Can I set the above configuration?")) {
        listRewind(cluster_manager.nodes, &li);
        while ((ln = listNext(&li)) != NULL) {
            clusterManagerNode *node = ln->value;
            char *err = NULL;
            int flushed = clusterManagerFlushNodeConfig(node, &err);
            if (!flushed && node->dirty && !node->replicate) {
                if (err != NULL) {
                    CLUSTER_MANAGER_PRINT_REPLY_ERROR(node, err);
                    zfree(err);
                }
                success = 0;
                goto cleanup;
            } else if (err != NULL) zfree(err);
        }
        clusterManagerLogInfo(">>> Nodes configuration updated\n");
        clusterManagerLogInfo(">>> Assign a different config epoch to "

And I don't get how confirmWithYes can use the --cluster-yet.

static int confirmWithYes(char *msg) {
    printf("%s (type 'yes' to accept): ", msg);
    fflush(stdout);
    char buf[4];
    int nread = read(fileno(stdin),buf,4);
    buf[3] = '\0';
    return (nread != 0 && !strcmp("yes", buf));
}

It seems to only be used during resharding

        } else if (!strcmp(argv[i],"--cluster-yes")) {
            config.cluster_manager_command.flags |=
                CLUSTER_MANAGER_CMD_FLAG_YES;

...

    if (!(config.cluster_manager_command.flags &
          CLUSTER_MANAGER_CMD_FLAG_YES))
    {
        printf("Do you want to proceed with the proposed "
               "reshard plan (yes/no)? ");
        fflush(stdout);
        char buf[4];
        int nread = read(fileno(stdin),buf,4);
        buf[3] = '\0';
        if (nread <= 0 || strcmp("yes", buf) != 0) {
            result = 0;
            goto cleanup;
        }
    }

... I believe this patch should fix it, I'll give it a try:

$ git diff
diff --git a/src/redis-cli.c b/src/redis-cli.c
index 208f7b35d..53b794fbf 100644
--- a/src/redis-cli.c
+++ b/src/redis-cli.c
@@ -5511,7 +5511,9 @@ assign_replicas:
     }
     clusterManagerOptimizeAntiAffinity(ip_nodes, ip_count);
     clusterManagerShowNodes();
-    if (confirmWithYes("Can I set the above configuration?")) {
+
+    if ((config.cluster_manager_command.flags & CLUSTER_MANAGER_CMD_FLAG_YES) ||
+        confirmWithYes("Can I set the above configuration?")) {
         listRewind(cluster_manager.nodes, &li);
         while ((ln = listNext(&li)) != NULL) {
             clusterManagerNode *node = ln->value;

ps: ... shameless plug advertisement for my python tool to do similar task ;p / https://github.com/machinezone/rcc/ (which has a resharding and a manual cluster creation mode for redis-4)

Comment From: miguelaeh

Hi @bsergean , Thank you very much for the quick PR!!!

Comment From: bsergean

Sure thing !

We'll see what Salvatore says but I think that the PR or a variation of it should be merged eventually.

On May 14, 2020, at 12:55 AM, Miguel Ángel Cabrera Miñagorri notifications@github.com wrote:

Hi @bsergean https://github.com/bsergean , Thank you very much for the quick PR!!!

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/antirez/redis/issues/7246#issuecomment-628461117, or unsubscribe https://github.com/notifications/unsubscribe-auth/AC2O6UPFSEYSCD4IZX7CHNLRROPWRANCNFSM4M7UEBPA.