I’ve tried rolling upgrade, but stacked in rolling-restart.sh point.
What I did is,
1. Installed a new version 1.1.2
2. Change configs(600xx ports instead of 160xx)
3. Change the link to new installation
4. Run ./rolling-restart.sh —config ./conf from new installation directory
After this, I got
stopping master.
Waiting for Master ZNode /hbase/master to expire……………………………………………………
I’ve checked the zookeeper logs, and they’re full of the following messages
2015-11-09 20:53:54,673 [myid:] - INFO
[NIOServerCxn.Factory:0.0.0.0/0.0.0.0:2181:NIOServerCnxnFactory@197] - Accepted
socket connection from /192.168.56.160:51176
2015-11-09 20:53:54,675 [myid:] - INFO
[NIOServerCxn.Factory:0.0.0.0/0.0.0.0:2181:ZooKeeperServer@868] - Client
attempting to establish new session at /192.168.56.160:51176
2015-11-09 20:53:54,677 [myid:] - INFO [SyncThread:0:ZooKeeperServer@617] -
Established session 0x150ed71c0f902b3 with negotiated timeout 30000 for client
/192.168.56.160:51176
2015-11-09 20:53:55,013 [myid:] - WARN
[NIOServerCxn.Factory:0.0.0.0/0.0.0.0:2181:NIOServerCnxn@357] - caught end of
stream exception
EndOfStreamException: Unable to read additional data from client sessionid
0x150ed71c0f902b3, likely client has closed socket
at
org.apache.zookeeper.server.NIOServerCnxn.doIO(NIOServerCnxn.java:228)
at
org.apache.zookeeper.server.NIOServerCnxnFactory.run(NIOServerCnxnFactory.java:208)
at java.lang.Thread.run(Thread.java:745)
2015-11-09 20:53:55,015 [myid:] - INFO
[NIOServerCxn.Factory:0.0.0.0/0.0.0.0:2181:NIOServerCnxn@1007] - Closed socket
connection for client /192.168.56.160:51176 which had sessionid
0x150ed71c0f902b3
2015-11-09 20:53:56,707 [myid:] - INFO
[NIOServerCxn.Factory:0.0.0.0/0.0.0.0:2181:NIOServerCnxnFactory@197] - Accepted
socket connection from /192.168.56.160:51178
2015-11-09 20:53:56,709 [myid:] - INFO
[NIOServerCxn.Factory:0.0.0.0/0.0.0.0:2181:ZooKeeperServer@868] - Client
attempting to establish new session at /192.168.56.160:51178
2015-11-09 20:53:56,712 [myid:] - INFO [SyncThread:0:ZooKeeperServer@617] -
Established session 0x150ed71c0f902b4 with negotiated timeout 30000 for client
/192.168.56.160:51178
2015-11-09 20:53:57,048 [myid:] - WARN
[NIOServerCxn.Factory:0.0.0.0/0.0.0.0:2181:NIOServerCnxn@357] - caught end of
stream exception
EndOfStreamException: Unable to read additional data from client sessionid
0x150ed71c0f902b4, likely client has closed socket
at
org.apache.zookeeper.server.NIOServerCnxn.doIO(NIOServerCnxn.java:228)
at
org.apache.zookeeper.server.NIOServerCnxnFactory.run(NIOServerCnxnFactory.java:208)
at java.lang.Thread.run(Thread.java:745)
2015-11-09 20:53:57,051 [myid:] - INFO
[NIOServerCxn.Factory:0.0.0.0/0.0.0.0:2181:NIOServerCnxn@1007] - Closed socket
connection for client /192.168.56.160:51178 which had sessionid
0x150ed71c0f902b4
Then I tried without rolling upgrade, the steps were
1. Install new version
2. Change configs(600xx ports instead of 160xx)
3. Change the link to new installation
4. Stop all daemons
5. Start all daemons
I’ve checked hbck and the state is consistent. Is it valid way of upgrading an
hbase, I know there is a downtime, but I can tolerate this.
Thank you.
> On 09 Nov 2015, at 15:23, Ted Yu <[email protected]> wrote:
>
> The description covers 0.98.x to 1.1.y as well.
>
> Rolling upgrade is preferred because there is no downtime.
>
> Cheers
>
>> On Nov 9, 2015, at 2:53 AM, Akmal Abbasov <[email protected]> wrote:
>>
>> Hi Ted,
>> thank you for your reply.
>> The link you’ve provided contains information for 0.98.x to 1.0.x, but does
>> it apply for 0.98.x to 1.1.x as well?
>> Moreover I found that only rolling upgrade is discussed, what about the
>> upgrade with turning off the cluster.
>> Thank you.
>>
>>
>>> On 09 Nov 2015, at 09:38, Ted Yu <[email protected]> wrote:
>>>
>>> Please take a look at
>>> http://hbase.apache.org/book.html#_upgrade_paths
>>>
>>>> On Nov 8, 2015, at 11:47 PM, Akmal Abbasov <[email protected]>
>>>> wrote:
>>>>
>>>> Hi all,
>>>> I’m planing to upgrade my HBase. Currently it’s hbase-0.98.7-hadoop2. I
>>>> was wondering can I upgrade directly to current stable version, which is
>>>> 1.1.2?
>>>> Moreover, I can tolerate downtime, so in this case upgrade process will
>>>> include,
>>>> a. Installation of the new version
>>>> b. Distribution of configurations
>>>> c. Change links to reference to a new version
>>>> d. $HBASE_HOME/bin/hbase upgrade
>>>> e. Start upgraded cluster
>>>> Is this the all required steps which should be done?
>>>> Thank you.
>>
