You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@doris.apache.org by GitBox <gi...@apache.org> on 2019/09/13 05:37:02 UTC
[GitHub] [incubator-doris] kangkaisen opened a new issue #1801: Critical
bug: Cluster meta will write wrong!
kangkaisen opened a new issue #1801: Critical bug: Cluster meta will write wrong!
URL: https://github.com/apache/incubator-doris/issues/1801
**Describe the bug**
Today, I found our Prod FE metadata was broken again!!!
```
2019-09-13 10:21:18,027 WARN 1391 [Cluster.addDb():136] addDb xiaoxiang:information_schema 20586
2019-09-13 10:21:18,027 WARN 1391 [Cluster.readFields():258] dbNames is [caiwu:finance_mart]
2019-09-13 10:21:18,027 WARN 1391 [Cluster.readFields():263] db id is 10000
2019-09-13 10:21:18,027 WARN 1391 [Cluster.readFields():268] linkDbNames count 0
2019-09-13 10:21:18,027 WARN 1391 [Cluster.readFields():275] read linkDbNames {}
2019-09-13 10:21:18,027 WARN 1391 [Cluster.readFields():278] linkDbIds count 55506441
2019-09-13 10:21:19,853 ERROR 1391 [Checkpoint.runOneCycle():123] Exception when generate new image file
java.io.EOFException: null
at java.io.DataInputStream.readFully(DataInputStream.java:202) ~[?:?]
at org.apache.doris.common.io.Text.readString(Text.java:416) ~[palo-fe.jar:?]
at org.apache.doris.persist.LinkDbInfo.readFields(LinkDbInfo.java:69) ~[palo-fe.jar:?]
at org.apache.doris.cluster.Cluster.readFields(Cluster.java:282) ~[palo-fe.jar:?]
```
```
2019-09-11 00:07:18,861 WARN 1391 [Cluster.addDb():136] addDb default_cluster:information_schema 9403
2019-09-10 00:23:44,081 WARN 1391 [Cluster.addDb():136] addDb default_cluster:information_schema 8773
```
When FE master run a long time, if you use Cluster feature, the Checkpoint will fail!
The logic chain is:
1 Every time we `loadCluster`, we will new `InfoSchemaDb`.
```
final InfoSchemaDb db = new InfoSchemaDb(cluster.getName());
db.setClusterName(cluster.getName());
```
2 The InfoSchemaDb id will become bigger and bigger.
3 When InfoSchemaDb id bigger than `Catalog.NEXT_ID_INIT_VALUE`, the Cluster meta will write wrong!
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
With regards,
Apache Git Services
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@doris.apache.org
For additional commands, e-mail: dev-help@doris.apache.org