info-cmd: add list of resize failed nodes

[pkalever]

What does this PR achieve? Why do we need it?

$ gluster-block modify blockhostingvol/blockvol size 4GiB --json-pretty
{
  "IQN":"iqn.2016-12.org.gluster-block:7a4938b1-f153-4de7-b91a-04c34fb57f36",
  "FAILED ON":[
    "192.168.121.55",
    "192.168.121.206"
  ],
  "SUCCESSFUL ON":[
    "192.168.121.253"
  ],
  "RESULT":"FAIL"
}

$ gluster-block info blockhostingvol/blockvol
NAME: blockvol
VOLUME: blockhostingvol
GBID: 7a4938b1-f153-4de7-b91a-04c34fb57f36
SIZE: 4.0 GiB
HA: 3
IOTIMEOUT: 0 Seconds
PASSWORD:
EXPORTED ON: 192.168.121.253 192.168.121.55 192.168.121.206
RESIZE FAILED ON: 192.168.121.55:[2.0 GiB] 192.168.121.206:[1.0 GiB]

$ gluster-block info blockhostingvol/blockvol --json-pretty
{
  "NAME":"blockvol",
  "VOLUME":"blockhostingvol",
  "GBID":"7a4938b1-f153-4de7-b91a-04c34fb57f36",
  "SIZE":"4.0 GiB",
  "HA":3,
  "IOTIMEOUT":"0 Seconds",
  "PASSWORD":"",
  "EXPORTED ON":[
    "192.168.121.253",
    "192.168.121.55",
    "192.168.121.206"
  ],
  "RESIZE FAILED ON":{
    "192.168.121.55":"2.0 GiB",
    "192.168.121.206":"1.0 GiB"
  }
}

Notes for the reviewer

This is needed by heketi for expand future integration.

Signed-off-by: Prasanna Kumar Kalever \prasanna.kalever@redhat.com\

[pkalever]

@phlogistonjohn @raghavendra-talur FYI. Just in case if you want to see any additional details.

[lxbsz]

@pkalever BTW, does add the old size for failing resize nodes make sense here ?

[pkalever]

@pkalever BTW, does add the old size for failing resize nodes make sense here ?

@lxbsz make sense, done now.

Thanks!

[pkalever]

@lxbsz Thanks for testing this, let's wait for a day on this, I would like to hear back from @phlogistonjohn @raghavendra-talur

[raghavendra-talur]

I have not looked at the code.

The data in the output looks good and sufficient. However it does not seem like "RESIZE FAILED ON" fits in the info structure. Taking inspiration from the kuberentes PV info yaml, I think if you keep a status field in the output then you can have a value something along the lines of needs resizing and then a per node message.

"status": {
       "message": {
           "10.72.36.245": "14.0 GiB"
        }
        "phase": "Needs Resize"
}

Only take that suggestion if it is doable and if it also applies well to other workflows in gluster-block, like create, delete etc.

[pkalever]

Only take that suggestion if it is doable and if it also applies well to other workflows in gluster-block, like create, delete etc.

@raghavendra-talur Thanks! I went back and checked the code, and this is not only the place using such kind of naming, its in gluster-block genes :-)

rpc/block_create.c:                                    "ROLLBACK FAILED ON");
rpc/block_create.c:      if (GB_ASPRINTF(&tmp, "ROLLBACK FAILED ON: %s\n",
rpc/block_info.c:      json_object_object_add(json_obj, "RESIZE FAILED ON", json_obj2);
rpc/block_info.c:      if (GB_ASPRINTF(&tmp, "%s\nRESIZE FAILED ON:%s", tmp4, rsf_nodes) == -1) {
rpc/block_modify.c:                                   "ROLLBACK FAILED ON");
rpc/block_modify.c:        GB_ASPRINTF(&tmp, "ROLLBACK FAILED ON: %s\n", savereply->rb_attempt);
rpc/block_replace.c:                                     "REPLACE PORTAL FAILED ON");
rpc/block_replace.c:          if (GB_ASPRINTF(&entry, "%sREPLACE PORTAL FAILED ON: %s\n", tmp?tmp:"",

Hence for consistency purpose many be we should continue using it :-)

[pkalever]

@lxbsz refreshed with the changes discussed, please take a look. Thanks!

[pkalever]

Thanks @lxbsz, merging now!