Details
Description
After master-RS colocation, now the pseudo dist-mode does not work as documented since you cannot start a region server in the same port 16020.
I think we can either select a random port (and info port) for the master's region server, or document how to do a pseudo-distributed setup in the book.
Jimmy Xiang wdyt?
Issue Links
- is duplicated by
-
HBASE-11669
The HMaster uses the same port with HRegionServer
-
- Resolved
-
- is related to
-
HBASE-11604
Disable co-locating meta/master by default
-
- Closed
-
Activity
| Field | Original Value | New Value |
|---|---|---|
| Assignee | Jimmy Xiang [ jxiang ] |
| Attachment | hbase-11575.patch [ 12657432 ] |
| Status | Open [ 1 ] | Patch Available [ 10002 ] |
Looked into it and attached a patch that fixed some documentation mismatches, changed the pseudo distributed mode setting a little so that we just start one instance. Additional regionservers can be started with the local regionserver sh script.
So we are not doing the random port approach? It should be ok.
With hbase.master.wait.on.regionservers.mintostart set to 1, master should be able to finish initialization. Can user regions be opened in this mode from the RS inside msater?
We can do random port for web UI, but not for IPC port. Yes, user regions can be opened in the RS inside the master. If additional regionservers are started, balancer can move user regions out of the master.
We probably should avoid random port.
See HBASE-10289.
1 for branch-1 if you have tested the script changes. Thanks Jimmy for nailing this.
Yes, the patch worked for me. Pseudo distributed mode is fixed. Thanks for reporting the issue and reviewing the patch. Integrated in branch 1 and master.
| Status | Patch Available [ 10002 ] | Resolved [ 5 ] |
| Hadoop Flags | Reviewed [ 10343 ] | |
| Fix Version/s | 1.0.0 [ 12325852 ] | |
| Fix Version/s | 2.0.0 [ 12327188 ] | |
| Fix Version/s | 0.99.0 [ 12325675 ] | |
| Resolution | Fixed [ 1 ] |
SUCCESS: Integrated in HBase-TRUNK #5340 (See https://builds.apache.org/job/HBase-TRUNK/5340/)
HBASE-11575 Pseudo distributed mode does not work as documented (jxiang: rev cc61cc308168af407a2f21851e9932292af8ca77)
- conf/regionservers
- bin/hbase-config.sh
- bin/local-master-backup.sh
- bin/local-regionservers.sh
- src/main/docbkx/getting_started.xml
SUCCESS: Integrated in HBase-1.0 #68 (See https://builds.apache.org/job/HBase-1.0/68/)
HBASE-11575 Pseudo distributed mode does not work as documented (jxiang: rev 147a3521f94f82c3a189a4ace2f7232c48b1ad5d)
- bin/local-master-backup.sh
- bin/local-regionservers.sh
- bin/hbase-config.sh
- conf/regionservers
- src/main/docbkx/getting_started.xml
| Release Note | To start a cluster in pseudo distributed mode, in addition to set hbase.cluster.distributed to true, hbase.master.wait.on.regionservers.mintostart should be set to 1. |
| Link |
This issue is blocked by |
| Link |
This issue is blocked by |
| Link |
This issue is duplicated by |
hbase.master.wait.on.regionservers.mintostart should be set to 1.
Does it have to be this way Jimmy? This is in branch-1 too? That'd be a pain for users. Would be ok for master I'd say.
Unfortunately, it is in branch-1 too. The problem is that master uses the same RPC port as a region server now. Not sure how to handle this better.
| Fix Version/s | 0.99.0 [ 12325675 ] | |
| Fix Version/s | 1.0.0 [ 12325852 ] |
| Link |
This issue is related to |
| Release Note | To start a cluster in pseudo distributed mode, in addition to set hbase.cluster.distributed to true, hbase.master.wait.on.regionservers.mintostart should be set to 1. |
Release note is removed since we are going to change the code so that additional config is not needed. See HBASE-11604.
Closing this issue after 0.99.0 release.
| Status | Resolved [ 5 ] | Closed [ 6 ] |
Probably it's better to use a random port. Let me take a look.