Github user asfgit closed the pull request at:
https://github.com/apache/flink/pull/1227
---
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 enab
Github user tillrohrmann commented on the pull request:
https://github.com/apache/flink/pull/1227#issuecomment-146954580
Looks good. I will rebase it on the latest state recovery branch and if
travis gives green light, then I'll merge it.
---
If your project is set up for it, you can
Github user uce commented on the pull request:
https://github.com/apache/flink/pull/1227#issuecomment-145876037
- It is very tightly coupled in in case of recovery. We won't loose this,
but I agree to introduce an interface to hide this (with a no-op implementation
for standalone mode
Github user tillrohrmann commented on a diff in the pull request:
https://github.com/apache/flink/pull/1227#discussion_r41270908
--- Diff:
flink-runtime/src/main/java/org/apache/flink/runtime/blob/BlobServer.java ---
@@ -70,18 +82,65 @@
private final int maxConnections;
Github user uce commented on a diff in the pull request:
https://github.com/apache/flink/pull/1227#discussion_r41269987
--- Diff:
flink-runtime/src/main/java/org/apache/flink/runtime/blob/BlobServer.java ---
@@ -70,18 +82,65 @@
private final int maxConnections;
Github user tillrohrmann commented on the pull request:
https://github.com/apache/flink/pull/1227#issuecomment-145812243
I think it makes a difference if the JVM is killed with SIGTERM or SIGKILL.
In the former case, the shutdown hooks will be executed. I might be that Yarn
first send
Github user tillrohrmann commented on the pull request:
https://github.com/apache/flink/pull/1227#issuecomment-145810928
Good PR @uce. I think the proposal in your description to let the
client-side (TMs) 1. check whether they have the file cached, 2. check the
filesystem backend and
Github user uce commented on a diff in the pull request:
https://github.com/apache/flink/pull/1227#discussion_r41246346
--- Diff:
flink-runtime/src/main/java/org/apache/flink/runtime/blob/BlobServer.java ---
@@ -70,18 +82,65 @@
private final int maxConnections;
Github user uce commented on the pull request:
https://github.com/apache/flink/pull/1227#issuecomment-145808965
Exactly. If the blob server dies, then the job manager will still shutdown
the cache service though. This is the same behaviour as in the job manager
recovery. Thinking abou
Github user uce commented on a diff in the pull request:
https://github.com/apache/flink/pull/1227#discussion_r41246319
--- Diff:
flink-runtime/src/test/java/org/apache/flink/runtime/blob/BlobRecoveryITCase.java
---
@@ -0,0 +1,159 @@
+/*
+ * Licensed to the Apache Software
Github user uce commented on a diff in the pull request:
https://github.com/apache/flink/pull/1227#discussion_r41245655
--- Diff:
flink-runtime/src/main/java/org/apache/flink/runtime/blob/BlobServer.java ---
@@ -102,8 +161,13 @@ public BlobServer(Configuration config) throws
IOExc
Github user tillrohrmann commented on the pull request:
https://github.com/apache/flink/pull/1227#issuecomment-145799064
What happens with the jars on HDFS if the last `BlobServer` dies without
properly calling `shutdown`? They will only be removed if a new Flink cluster
is started wi
Github user tillrohrmann commented on a diff in the pull request:
https://github.com/apache/flink/pull/1227#discussion_r41243069
--- Diff:
flink-runtime/src/main/java/org/apache/flink/runtime/blob/BlobServer.java ---
@@ -102,8 +161,13 @@ public BlobServer(Configuration config) thro
Github user tillrohrmann commented on a diff in the pull request:
https://github.com/apache/flink/pull/1227#discussion_r41242183
--- Diff:
flink-runtime/src/test/java/org/apache/flink/runtime/execution/librarycache/BlobLibraryCacheRecoveryITCase.java
---
@@ -0,0 +1,174 @@
+/*
Github user tillrohrmann commented on a diff in the pull request:
https://github.com/apache/flink/pull/1227#discussion_r41241923
--- Diff:
flink-runtime/src/test/java/org/apache/flink/runtime/blob/BlobRecoveryITCase.java
---
@@ -0,0 +1,159 @@
+/*
+ * Licensed to the Apache
Github user tillrohrmann commented on a diff in the pull request:
https://github.com/apache/flink/pull/1227#discussion_r41239095
--- Diff:
flink-runtime/src/main/java/org/apache/flink/runtime/blob/BlobServer.java ---
@@ -102,8 +161,13 @@ public BlobServer(Configuration config) thro
Github user tillrohrmann commented on a diff in the pull request:
https://github.com/apache/flink/pull/1227#discussion_r41238986
--- Diff:
flink-runtime/src/main/java/org/apache/flink/runtime/blob/BlobServer.java ---
@@ -70,18 +82,65 @@
private final int maxConnections;
Github user tillrohrmann commented on a diff in the pull request:
https://github.com/apache/flink/pull/1227#discussion_r41238757
--- Diff:
flink-runtime/src/main/java/org/apache/flink/runtime/blob/BlobServer.java ---
@@ -70,18 +82,65 @@
private final int maxConnections;
GitHub user uce opened a pull request:
https://github.com/apache/flink/pull/1227
[FLINK-2805] [blobmanager] Write JARs to file state backend for recovery
This is a follow up to #1153. I've taken two changes from #1153 for
convenience. Other than that, this is independent.
W
19 matches
Mail list logo
