Github user revans2 commented on a diff in the pull request:
https://github.com/apache/storm/pull/2113#discussion_r120658098
--- Diff:
storm-server/src/main/java/org/apache/storm/localizer/AsyncLocalizer.java ---
@@ -144,8 +148,19 @@ public Void call() throws Exception {
File tr = new File(tmproot);
try {
downloadBaseBlobs(tr);
+ if (_assignment.is_set_total_node_shared()) {
+ File sharedMemoryDirTmpLocation = new File(tr,
"shared_by_topology");
+ //We need to create a directory for shared memory
to write to (we should not encourage this though)
+ Path path = sharedMemoryDirTmpLocation.toPath();
+ Files.createDirectories(path);
+ }
_fsOps.moveDirectoryPreferAtomic(tr, _stormRoot);
-
_fsOps.setupStormCodeDir(ConfigUtils.readSupervisorStormConf(_conf,
_topologyId), _stormRoot);
+ Map<String, Object> topoConf =
ConfigUtils.readSupervisorStormConf(_conf, _topologyId);
+ _fsOps.setupStormCodeDir(topoConf, _stormRoot);
+ if (_assignment.is_has_node_shared_memory()) {
+ File sharedMemoryDir = new File(_stormRoot,
"shared_by_topology");
--- End diff --
That is up to the topology. It could be JNI code or even something that a
ShellBolt or ShellSpout does. The point is to provide a place for them to
store the data if they need it. We have a group doing this in /tmp right now
and causing all kinds of issues if things crash/get rescheduled leaking things.
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at [email protected] or file a JIRA ticket
with INFRA.
---
