You are viewing a plain text version of this content. The canonical link for it is here.
Posted to user@hbase.apache.org by Ted Yu <yu...@gmail.com> on 2015/11/18 19:47:57 UTC
Re: HBase client and server version is not compatible
leadregionserver down
HBaseAdmin has this method:
public ClusterStatus getClusterStatus() throws IOException {
where ClusterStatus has:
public String getHBaseVersion() {
return hbaseVersion;
FYI
On Wed, Nov 18, 2015 at 8:58 AM, 聪聪 <17...@qq.com> wrote:
> Ted：
>
>
> Because a lot of development is the use of the wrong client, I think about
> how to avoid.For example, we even upgrade to 1.0 but they may use 2.0
> version.
>
>
>
>
> ------------------ 原始邮件 ------------------
> 发件人: "Ted Yu";<yu...@gmail.com>;
> 发送时间: 2015年11月18日(星期三) 晚上7:37
> 收件人: "user@hbase.apache.org"<us...@hbase.apache.org>;
>
> 主题: Re: HBase client and server version is not compatible leadregionserver
> down
>
>
>
> See http://hbase.apache.org/book.html#hbase.rolling.upgrade
>
> For example, in Rolling upgrade from 0.98.x to HBase 1.0.0, we state that
> it is possible to do a rolling upgrade between hbase-0.98.x and
> hbase-1.0.0.
>
> Cheers
>
> On Wed, Nov 18, 2015 at 12:22 AM, 聪聪 <17...@qq.com> wrote:
>
> > We recently found that regionserver down.Later, we found that because the
> > client and server version is not compatible.The client version is
> > 1.0,server version is 0.98.6.I want to know why this is, and whether
> there
> > is a better protection mechanism.How to avoid this problem, because some
> > development will appear this kind of mistake operation.
>
Re: HBase client and server version is not compatible
leadregionserver down
Posted by "Vincent.Gong" <th...@hotmail.com>.
To Ted,
I have no time now , I will reproduce the scene next week ,and
feedback the log for u.
regards,
many thanks.
在 2015年11月19日 21:45, Ted Yu 写道:
> Vincent:
> Can you pastebin region server log around the time of crash ?
>
> Cheers
>
>> On Nov 19, 2015, at 12:15 AM, Vincent.Gong <th...@hotmail.com> wrote:
>>
>> Hi All,
>> I have also found a similar problem.
>> If the regionserver and client's version was not match, the regionserver would crash .
>>
>> the crash case: regionserver's version is 1.0 and client's version is 1.1.0.
>>
>>
>>
>>> 在 2015年11月19日 14:53, Yu Li 写道:
>>> Second Enis and Heng, suggest to dig deeper into RS log and find the real
>>> reason causing RS crash.
>>>
>>> Best Regards,
>>> Yu
>>>
>>>> On 19 November 2015 at 03:18, Enis Söztutar <en...@apache.org> wrote:
>>>>
>>>> It is highly unlikely that a client to cause the server side to abort. It
>>>> is possible that due to some other problem the regionservers are aborting.
>>>> The regionservers will reject client connection requests if there is an RPC
>>>> version mismatch.
>>>>
>>>> 1.x and 0.98 client and servers have been tested to be rolling upgrade
>>>> compatible (meaning that older clients can work with newer server versions
>>>> or vice versa).
>>>>
>>>> Enis
>>>>
>>>>> On Wed, Nov 18, 2015 at 10:47 AM, Ted Yu <yu...@gmail.com> wrote:
>>>>>
>>>>> HBaseAdmin has this method:
>>>>>
>>>>> public ClusterStatus getClusterStatus() throws IOException {
>>>>>
>>>>> where ClusterStatus has:
>>>>>
>>>>> public String getHBaseVersion() {
>>>>>
>>>>> return hbaseVersion;
>>>>>
>>>>> FYI
>>>>>
>>>>>> On Wed, Nov 18, 2015 at 8:58 AM, 聪聪 <17...@qq.com> wrote:
>>>>>>
>>>>>> Ted：
>>>>>>
>>>>>>
>>>>>> Because a lot of development is the use of the wrong client, I think
>>>>> about
>>>>>> how to avoid.For example, we even upgrade to 1.0 but they may use 2.0
>>>>>> version.
>>>>>>
>>>>>>
>>>>>>
>>>>>>
>>>>>> ------------------ 原始邮件 ------------------
>>>>>> 发件人: "Ted Yu";<yu...@gmail.com>;
>>>>>> 发送时间: 2015年11月18日(星期三) 晚上7:37
>>>>>> 收件人: "user@hbase.apache.org"<us...@hbase.apache.org>;
>>>>>>
>>>>>> 主题: Re: HBase client and server version is not compatible
>>>>> leadregionserver
>>>>>> down
>>>>>>
>>>>>>
>>>>>>
>>>>>> See http://hbase.apache.org/book.html#hbase.rolling.upgrade
>>>>>>
>>>>>> For example, in Rolling upgrade from 0.98.x to HBase 1.0.0, we state
>>>> that
>>>>>> it is possible to do a rolling upgrade between hbase-0.98.x and
>>>>>> hbase-1.0.0.
>>>>>>
>>>>>> Cheers
>>>>>>
>>>>>>> On Wed, Nov 18, 2015 at 12:22 AM, 聪聪 <17...@qq.com> wrote:
>>>>>>>
>>>>>>> We recently found that regionserver down.Later, we found that because
>>>>> the
>>>>>>> client and server version is not compatible.The client version is
>>>>>>> 1.0,server version is 0.98.6.I want to know why this is, and whether
>>>>>> there
>>>>>>> is a better protection mechanism.How to avoid this problem, because
>>>>> some
>>>>>>> development will appear this kind of mistake operation.
>
Re: HBase client and server version is not compatible leadregionserver down
Posted by Ted Yu <yu...@gmail.com>.
Vincent:
Can you pastebin region server log around the time of crash ?
Cheers
> On Nov 19, 2015, at 12:15 AM, Vincent.Gong <th...@hotmail.com> wrote:
>
> Hi All,
> I have also found a similar problem.
> If the regionserver and client's version was not match, the regionserver would crash .
>
> the crash case: regionserver's version is 1.0 and client's version is 1.1.0.
>
>
>
>> 在 2015年11月19日 14:53, Yu Li 写道:
>> Second Enis and Heng, suggest to dig deeper into RS log and find the real
>> reason causing RS crash.
>>
>> Best Regards,
>> Yu
>>
>>> On 19 November 2015 at 03:18, Enis Söztutar <en...@apache.org> wrote:
>>>
>>> It is highly unlikely that a client to cause the server side to abort. It
>>> is possible that due to some other problem the regionservers are aborting.
>>> The regionservers will reject client connection requests if there is an RPC
>>> version mismatch.
>>>
>>> 1.x and 0.98 client and servers have been tested to be rolling upgrade
>>> compatible (meaning that older clients can work with newer server versions
>>> or vice versa).
>>>
>>> Enis
>>>
>>>> On Wed, Nov 18, 2015 at 10:47 AM, Ted Yu <yu...@gmail.com> wrote:
>>>>
>>>> HBaseAdmin has this method:
>>>>
>>>> public ClusterStatus getClusterStatus() throws IOException {
>>>>
>>>> where ClusterStatus has:
>>>>
>>>> public String getHBaseVersion() {
>>>>
>>>> return hbaseVersion;
>>>>
>>>> FYI
>>>>
>>>>> On Wed, Nov 18, 2015 at 8:58 AM, 聪聪 <17...@qq.com> wrote:
>>>>>
>>>>> Ted：
>>>>>
>>>>>
>>>>> Because a lot of development is the use of the wrong client, I think
>>>> about
>>>>> how to avoid.For example, we even upgrade to 1.0 but they may use 2.0
>>>>> version.
>>>>>
>>>>>
>>>>>
>>>>>
>>>>> ------------------ 原始邮件 ------------------
>>>>> 发件人: "Ted Yu";<yu...@gmail.com>;
>>>>> 发送时间: 2015年11月18日(星期三) 晚上7:37
>>>>> 收件人: "user@hbase.apache.org"<us...@hbase.apache.org>;
>>>>>
>>>>> 主题: Re: HBase client and server version is not compatible
>>>> leadregionserver
>>>>> down
>>>>>
>>>>>
>>>>>
>>>>> See http://hbase.apache.org/book.html#hbase.rolling.upgrade
>>>>>
>>>>> For example, in Rolling upgrade from 0.98.x to HBase 1.0.0, we state
>>> that
>>>>> it is possible to do a rolling upgrade between hbase-0.98.x and
>>>>> hbase-1.0.0.
>>>>>
>>>>> Cheers
>>>>>
>>>>>> On Wed, Nov 18, 2015 at 12:22 AM, 聪聪 <17...@qq.com> wrote:
>>>>>>
>>>>>> We recently found that regionserver down.Later, we found that because
>>>> the
>>>>>> client and server version is not compatible.The client version is
>>>>>> 1.0,server version is 0.98.6.I want to know why this is, and whether
>>>>> there
>>>>>> is a better protection mechanism.How to avoid this problem, because
>>>> some
>>>>>> development will appear this kind of mistake operation.
>
Re: HBase client and server version is not compatible
leadregionserver down
Posted by "Vincent.Gong" <th...@hotmail.com>.
Hi All,
I have also found a similar problem.
If the regionserver and client's version was not match, the regionserver
would crash .
the crash case: regionserver's version is 1.0 and client's version is 1.1.0.
在 2015年11月19日 14:53, Yu Li 写道:
> Second Enis and Heng, suggest to dig deeper into RS log and find the real
> reason causing RS crash.
>
> Best Regards,
> Yu
>
> On 19 November 2015 at 03:18, Enis Söztutar <en...@apache.org> wrote:
>
>> It is highly unlikely that a client to cause the server side to abort. It
>> is possible that due to some other problem the regionservers are aborting.
>> The regionservers will reject client connection requests if there is an RPC
>> version mismatch.
>>
>> 1.x and 0.98 client and servers have been tested to be rolling upgrade
>> compatible (meaning that older clients can work with newer server versions
>> or vice versa).
>>
>> Enis
>>
>> On Wed, Nov 18, 2015 at 10:47 AM, Ted Yu <yu...@gmail.com> wrote:
>>
>>> HBaseAdmin has this method:
>>>
>>> public ClusterStatus getClusterStatus() throws IOException {
>>>
>>> where ClusterStatus has:
>>>
>>> public String getHBaseVersion() {
>>>
>>> return hbaseVersion;
>>>
>>> FYI
>>>
>>> On Wed, Nov 18, 2015 at 8:58 AM, 聪聪 <17...@qq.com> wrote:
>>>
>>>> Ted：
>>>>
>>>>
>>>> Because a lot of development is the use of the wrong client, I think
>>> about
>>>> how to avoid.For example, we even upgrade to 1.0 but they may use 2.0
>>>> version.
>>>>
>>>>
>>>>
>>>>
>>>> ------------------ 原始邮件 ------------------
>>>> 发件人: "Ted Yu";<yu...@gmail.com>;
>>>> 发送时间: 2015年11月18日(星期三) 晚上7:37
>>>> 收件人: "user@hbase.apache.org"<us...@hbase.apache.org>;
>>>>
>>>> 主题: Re: HBase client and server version is not compatible
>>> leadregionserver
>>>> down
>>>>
>>>>
>>>>
>>>> See http://hbase.apache.org/book.html#hbase.rolling.upgrade
>>>>
>>>> For example, in Rolling upgrade from 0.98.x to HBase 1.0.0, we state
>> that
>>>> it is possible to do a rolling upgrade between hbase-0.98.x and
>>>> hbase-1.0.0.
>>>>
>>>> Cheers
>>>>
>>>> On Wed, Nov 18, 2015 at 12:22 AM, 聪聪 <17...@qq.com> wrote:
>>>>
>>>>> We recently found that regionserver down.Later, we found that because
>>> the
>>>>> client and server version is not compatible.The client version is
>>>>> 1.0,server version is 0.98.6.I want to know why this is, and whether
>>>> there
>>>>> is a better protection mechanism.How to avoid this problem, because
>>> some
>>>>> development will appear this kind of mistake operation.
Re: HBase client and server version is not compatible
leadregionserver down
Posted by Yu Li <ca...@gmail.com>.
Second Enis and Heng, suggest to dig deeper into RS log and find the real
reason causing RS crash.
Best Regards,
Yu
On 19 November 2015 at 03:18, Enis Söztutar <en...@apache.org> wrote:
> It is highly unlikely that a client to cause the server side to abort. It
> is possible that due to some other problem the regionservers are aborting.
> The regionservers will reject client connection requests if there is an RPC
> version mismatch.
>
> 1.x and 0.98 client and servers have been tested to be rolling upgrade
> compatible (meaning that older clients can work with newer server versions
> or vice versa).
>
> Enis
>
> On Wed, Nov 18, 2015 at 10:47 AM, Ted Yu <yu...@gmail.com> wrote:
>
> > HBaseAdmin has this method:
> >
> > public ClusterStatus getClusterStatus() throws IOException {
> >
> > where ClusterStatus has:
> >
> > public String getHBaseVersion() {
> >
> > return hbaseVersion;
> >
> > FYI
> >
> > On Wed, Nov 18, 2015 at 8:58 AM, 聪聪 <17...@qq.com> wrote:
> >
> > > Ted：
> > >
> > >
> > > Because a lot of development is the use of the wrong client, I think
> > about
> > > how to avoid.For example, we even upgrade to 1.0 but they may use 2.0
> > > version.
> > >
> > >
> > >
> > >
> > > ------------------ 原始邮件 ------------------
> > > 发件人: "Ted Yu";<yu...@gmail.com>;
> > > 发送时间: 2015年11月18日(星期三) 晚上7:37
> > > 收件人: "user@hbase.apache.org"<us...@hbase.apache.org>;
> > >
> > > 主题: Re: HBase client and server version is not compatible
> > leadregionserver
> > > down
> > >
> > >
> > >
> > > See http://hbase.apache.org/book.html#hbase.rolling.upgrade
> > >
> > > For example, in Rolling upgrade from 0.98.x to HBase 1.0.0, we state
> that
> > > it is possible to do a rolling upgrade between hbase-0.98.x and
> > > hbase-1.0.0.
> > >
> > > Cheers
> > >
> > > On Wed, Nov 18, 2015 at 12:22 AM, 聪聪 <17...@qq.com> wrote:
> > >
> > > > We recently found that regionserver down.Later, we found that because
> > the
> > > > client and server version is not compatible.The client version is
> > > > 1.0,server version is 0.98.6.I want to know why this is, and whether
> > > there
> > > > is a better protection mechanism.How to avoid this problem, because
> > some
> > > > development will appear this kind of mistake operation.
> > >
> >
>
Re: HBase client and server version is not compatible
leadregionserver down
Posted by Enis Söztutar <en...@apache.org>.
It is highly unlikely that a client to cause the server side to abort. It
is possible that due to some other problem the regionservers are aborting.
The regionservers will reject client connection requests if there is an RPC
version mismatch.
1.x and 0.98 client and servers have been tested to be rolling upgrade
compatible (meaning that older clients can work with newer server versions
or vice versa).
Enis
On Wed, Nov 18, 2015 at 10:47 AM, Ted Yu <yu...@gmail.com> wrote:
> HBaseAdmin has this method:
>
> public ClusterStatus getClusterStatus() throws IOException {
>
> where ClusterStatus has:
>
> public String getHBaseVersion() {
>
> return hbaseVersion;
>
> FYI
>
> On Wed, Nov 18, 2015 at 8:58 AM, 聪聪 <17...@qq.com> wrote:
>
> > Ted：
> >
> >
> > Because a lot of development is the use of the wrong client, I think
> about
> > how to avoid.For example, we even upgrade to 1.0 but they may use 2.0
> > version.
> >
> >
> >
> >
> > ------------------ 原始邮件 ------------------
> > 发件人: "Ted Yu";<yu...@gmail.com>;
> > 发送时间: 2015年11月18日(星期三) 晚上7:37
> > 收件人: "user@hbase.apache.org"<us...@hbase.apache.org>;
> >
> > 主题: Re: HBase client and server version is not compatible
> leadregionserver
> > down
> >
> >
> >
> > See http://hbase.apache.org/book.html#hbase.rolling.upgrade
> >
> > For example, in Rolling upgrade from 0.98.x to HBase 1.0.0, we state that
> > it is possible to do a rolling upgrade between hbase-0.98.x and
> > hbase-1.0.0.
> >
> > Cheers
> >
> > On Wed, Nov 18, 2015 at 12:22 AM, 聪聪 <17...@qq.com> wrote:
> >
> > > We recently found that regionserver down.Later, we found that because
> the
> > > client and server version is not compatible.The client version is
> > > 1.0,server version is 0.98.6.I want to know why this is, and whether
> > there
> > > is a better protection mechanism.How to avoid this problem, because
> some
> > > development will appear this kind of mistake operation.
> >
>