MySQL server has gone away

Hi,
We are seeing intermittent errors of the type:

failed to execute " SELECT * FROM cluster_info WHERE cluster =3D ?":=
MySQL server has gone away

The error "MySQL server has gone away" is the error from the db handle. Can=
anyone give me any pointers on why that happens? I looked up the documenta=
tion in MySQL docs and the most common reason seems to be that it happens i=
f the query size is very large or if there is a timeout. None of them seems=
to be a probable cause. The "max_allowed_packet" on the server is 16 MB an=
d as can be seen in the query above, the query is very small and nowhere ne=
ar the size limit. We also have a timeout setting (wait_timeout) of 10 minu=
tes and the above query for us cannot possibly take that amount of time. In=
any case, given the same query, it executes correctly 99% of time (so to s=
peak). It fails intermittently with the above error. What possibly could be=
the reason? I also looked at the max connections on the server at that tim=
e (around ~80) and it is much less than the limit we have (limit is 1000). =
How can I extract more information when this happens? This error message su=
cks since it does not tell me what exactly happened. The server version is =
5.1.45.

Thanks
Aveek=

--
MySQL General Mailing List
For list archives: http://lists.mysql.com/mysql
To unsubscribe: http://lists.mysql.com/mysql?unsub=3Dgcdmg-mysql-2@m.gmane.o rg

Re: MySQL server has gone away

> failed to execute " SELECT * FROM cluster_info WHERE cluster = ?":
> MySQL server has gone away
>
> The error "MySQL server has gone away" is the error from the db handle. Can
> anyone give me any pointers on why that happens? I looked up the
> documentation in MySQL docs and the most common reason seems to be that it
> happens if the query size is very large or if there is a timeout. None of
> them seems to be a probable cause. The "max_allowed_packet" on the server
> is 16 MB and as can be seen in the query above, the query is very small
> and nowhere near the size limit. We also have a timeout setting
> (wait_timeout) of 10 minutes and the above query for us cannot possibly
> take that amount of time. In any case, given the same query, it executes
> correctly 99% of time (so to speak). It fails intermittently with the
> above error. What possibly could be the reason? I also looked at the max
> connections on the server at that time (around ~80) and it is much less
> than the limit we have (limit is 1000). How can I extract more information
> when this happens? This error message sucks since it does not tell me what
> exactly happened. The server version is 5.1.45.

Can you access the error log of the server? That can probably shed more light
on the issue...
--
Rik Wasmus

--
MySQL General Mailing List
For list archives: http://lists.mysql.com/mysql
To unsubscribe: http://lists.mysql.com/mysql?unsub=gcdmg-mysql-2@m.gmane.org

Re: MySQL server has gone away

Nothing in the error log or the slow query log that suggests that the query=
size is too large or us taking too much time to execute.

Thanks
Aveek

On May 25, 2011, at 3:53 PM, Rik Wasmus wrote:

>> failed to execute " SELECT * FROM cluster_info WHERE cluster =3D =
?":
>> MySQL server has gone away
>>=20
>> The error "MySQL server has gone away" is the error from the db handle. =
Can
>> anyone give me any pointers on why that happens? I looked up the
>> documentation in MySQL docs and the most common reason seems to be that =
it
>> happens if the query size is very large or if there is a timeout. None o=
f
>> them seems to be a probable cause. The "max_allowed_packet" on the serve=
r
>> is 16 MB and as can be seen in the query above, the query is very small
>> and nowhere near the size limit. We also have a timeout setting
>> (wait_timeout) of 10 minutes and the above query for us cannot possibly
>> take that amount of time. In any case, given the same query, it executes
>> correctly 99% of time (so to speak). It fails intermittently with the
>> above error. What possibly could be the reason? I also looked at the max
>> connections on the server at that time (around ~80) and it is much less
>> than the limit we have (limit is 1000). How can I extract more informati=
on
>> when this happens? This error message sucks since it does not tell me wh=
at
>> exactly happened. The server version is 5.1.45.
>=20
> Can you access the error log of the server? That can probably shed more l=
ight=20
> on the issue...
> --=20
> Rik Wasmus
>=20
> --=20
> MySQL General Mailing List
> For list archives: http://lists.mysql.com/mysql
> To unsubscribe: http://lists.mysql.com/mysql?unsub=3Daveekm@yahoo-inc.=
com
>=20


--
MySQL General Mailing List
For list archives: http://lists.mysql.com/mysql
To unsubscribe: http://lists.mysql.com/mysql?unsub=3Dgcdmg-mysql-2@m.gmane.o rg

Re: MySQL server has gone away

--20cf3071cc3634771804a434272a
Content-Type: text/plain; charset=ISO-8859-1

I had experience with such type of error, It was due lack of resources
available to MySql, max connections exceeds on the server.
you can write a simple script which will grab and store output of 'show
processlist' every min. and later you cna investigate the issue.


On Wed, May 25, 2011 at 3:34 AM, Aveek Misra wrote:

> Nothing in the error log or the slow query log that suggests that the query
> size is too large or us taking too much time to execute.
>
> Thanks
> Aveek
>
> On May 25, 2011, at 3:53 PM, Rik Wasmus wrote:
>
> >> failed to execute " SELECT * FROM cluster_info WHERE cluster =
> ?":
> >> MySQL server has gone away
> >>
> >> The error "MySQL server has gone away" is the error from the db handle.
> Can
> >> anyone give me any pointers on why that happens? I looked up the
> >> documentation in MySQL docs and the most common reason seems to be that
> it
> >> happens if the query size is very large or if there is a timeout. None
> of
> >> them seems to be a probable cause. The "max_allowed_packet" on the
> server
> >> is 16 MB and as can be seen in the query above, the query is very small
> >> and nowhere near the size limit. We also have a timeout setting
> >> (wait_timeout) of 10 minutes and the above query for us cannot possibly
> >> take that amount of time. In any case, given the same query, it executes
> >> correctly 99% of time (so to speak). It fails intermittently with the
> >> above error. What possibly could be the reason? I also looked at the max
> >> connections on the server at that time (around ~80) and it is much less
> >> than the limit we have (limit is 1000). How can I extract more
> information
> >> when this happens? This error message sucks since it does not tell me
> what
> >> exactly happened. The server version is 5.1.45.
> >
> > Can you access the error log of the server? That can probably shed more
> light
> > on the issue...
> > --
> > Rik Wasmus
> >
> > --
> > MySQL General Mailing List
> > For list archives: http://lists.mysql.com/mysql
> > To unsubscribe:
> http://lists.mysql.com/mysql?unsub=aveekm@yahoo-inc.com
> >
>
>
> --
> MySQL General Mailing List
> For list archives: http://lists.mysql.com/mysql
> To unsubscribe:
> http://lists.mysql.com/mysql?unsub=aim.prabhat@gmail.com
>
>


--
Best Regards,

Prabhat Kumar
MySQL DBA

My Blog: http://adminlinux.blogspot.com
My LinkedIn: http://www.linkedin.com/in/profileprabhat

--20cf3071cc3634771804a434272a--

Re: MySQL server has gone away

--20cf301cc3c25162b904a43b4d1e
Content-Type: text/plain; charset=ISO-8859-1

'MySQL server has gone away'
Can be a network problem,
Just to increase complexity :)
On May 26, 2011 11:03 PM, "Prabhat Kumar" wrote:
> I had experience with such type of error, It was due lack of resources
> available to MySql, max connections exceeds on the server.
> you can write a simple script which will grab and store output of 'show
> processlist' every min. and later you cna investigate the issue.
>
>
> On Wed, May 25, 2011 at 3:34 AM, Aveek Misra wrote:
>
>> Nothing in the error log or the slow query log that suggests that the
query
>> size is too large or us taking too much time to execute.
>>
>> Thanks
>> Aveek
>>
>> On May 25, 2011, at 3:53 PM, Rik Wasmus wrote:
>>
>> >> failed to execute " SELECT * FROM cluster_info WHERE cluster =
>> ?":
>> >> MySQL server has gone away
>> >>
>> >> The error "MySQL server has gone away" is the error from the db
handle.
>> Can
>> >> anyone give me any pointers on why that happens? I looked up the
>> >> documentation in MySQL docs and the most common reason seems to be
that
>> it
>> >> happens if the query size is very large or if there is a timeout. None
>> of
>> >> them seems to be a probable cause. The "max_allowed_packet" on the
>> server
>> >> is 16 MB and as can be seen in the query above, the query is very
small
>> >> and nowhere near the size limit. We also have a timeout setting
>> >> (wait_timeout) of 10 minutes and the above query for us cannot
possibly
>> >> take that amount of time. In any case, given the same query, it
executes
>> >> correctly 99% of time (so to speak). It fails intermittently with the
>> >> above error. What possibly could be the reason? I also looked at the
max
>> >> connections on the server at that time (around ~80) and it is much
less
>> >> than the limit we have (limit is 1000). How can I extract more
>> information
>> >> when this happens? This error message sucks since it does not tell me
>> what
>> >> exactly happened. The server version is 5.1.45.
>> >
>> > Can you access the error log of the server? That can probably shed more
>> light
>> > on the issue...
>> > --
>> > Rik Wasmus
>> >
>> > --
>> > MySQL General Mailing List
>> > For list archives: http://lists.mysql.com/mysql
>> > To unsubscribe:
>> http://lists.mysql.com/mysql?unsub=aveekm@yahoo-inc.com
>> >
>>
>>
>> --
>> MySQL General Mailing List
>> For list archives: http://lists.mysql.com/mysql
>> To unsubscribe:
>> http://lists.mysql.com/mysql?unsub=aim.prabhat@gmail.com
>>
>>
>
>
> --
> Best Regards,
>
> Prabhat Kumar
> MySQL DBA
>
> My Blog: http://adminlinux.blogspot.com
> My LinkedIn: http://www.linkedin.com/in/profileprabhat

--20cf301cc3c25162b904a43b4d1e--

Re: MySQL server has gone away

--bcaec544eb6e204ad804a46db536
Content-Type: text/plain; charset=ISO-8859-1

Hi Aveek,
I would like to suggest some points here:
You could try increasing the max allowed packets to 128MB. Though you think
16MB is enough, increasing it is not going to affect the server. It is
dynamic value, so you could revert back the changes if you feel so.
Also make sure log_warnings=2 which will give more error messages in error
log.

This is the basic step for this error, if it is not working then we should
check for other options then.

Regards,
Vinodh.k



On Fri, May 27, 2011 at 11:05 AM, Claudio Nanni wrote:

> 'MySQL server has gone away'
> Can be a network problem,
> Just to increase complexity :)
> On May 26, 2011 11:03 PM, "Prabhat Kumar" wrote:
> > I had experience with such type of error, It was due lack of resources
> > available to MySql, max connections exceeds on the server.
> > you can write a simple script which will grab and store output of 'show
> > processlist' every min. and later you cna investigate the issue.
> >
> >
> > On Wed, May 25, 2011 at 3:34 AM, Aveek Misra
> wrote:
> >
> >> Nothing in the error log or the slow query log that suggests that the
> query
> >> size is too large or us taking too much time to execute.
> >>
> >> Thanks
> >> Aveek
> >>
> >> On May 25, 2011, at 3:53 PM, Rik Wasmus wrote:
> >>
> >> >> failed to execute " SELECT * FROM cluster_info WHERE cluster =
> >> ?":
> >> >> MySQL server has gone away
> >> >>
> >> >> The error "MySQL server has gone away" is the error from the db
> handle.
> >> Can
> >> >> anyone give me any pointers on why that happens? I looked up the
> >> >> documentation in MySQL docs and the most common reason seems to be
> that
> >> it
> >> >> happens if the query size is very large or if there is a timeout.
> None
> >> of
> >> >> them seems to be a probable cause. The "max_allowed_packet" on the
> >> server
> >> >> is 16 MB and as can be seen in the query above, the query is very
> small
> >> >> and nowhere near the size limit. We also have a timeout setting
> >> >> (wait_timeout) of 10 minutes and the above query for us cannot
> possibly
> >> >> take that amount of time. In any case, given the same query, it
> executes
> >> >> correctly 99% of time (so to speak). It fails intermittently with the
> >> >> above error. What possibly could be the reason? I also looked at the
> max
> >> >> connections on the server at that time (around ~80) and it is much
> less
> >> >> than the limit we have (limit is 1000). How can I extract more
> >> information
> >> >> when this happens? This error message sucks since it does not tell me
> >> what
> >> >> exactly happened. The server version is 5.1.45.
> >> >
> >> > Can you access the error log of the server? That can probably shed
> more
> >> light
> >> > on the issue...
> >> > --
> >> > Rik Wasmus
> >> >
> >> > --
> >> > MySQL General Mailing List
> >> > For list archives: http://lists.mysql.com/mysql
> >> > To unsubscribe:
> >> http://lists.mysql.com/mysql?unsub=aveekm@yahoo-inc.com
> >> >
> >>
> >>
> >> --
> >> MySQL General Mailing List
> >> For list archives: http://lists.mysql.com/mysql
> >> To unsubscribe:
> >> http://lists.mysql.com/mysql?unsub=aim.prabhat@gmail.com
> >>
> >>
> >
> >
> > --
> > Best Regards,
> >
> > Prabhat Kumar
> > MySQL DBA
> >
> > My Blog: http://adminlinux.blogspot.com
> > My LinkedIn: http://www.linkedin.com/in/profileprabhat
>

--bcaec544eb6e204ad804a46db536--

Re: MySQL server has gone away

Try setting wait_timeout=36000

Regards,
Chandru

On 5/29/2011 11:13 PM, ars k wrote:
> Hi Aveek,
> I would like to suggest some points here:
> You could try increasing the max allowed packets to 128MB. Though you think
> 16MB is enough, increasing it is not going to affect the server. It is
> dynamic value, so you could revert back the changes if you feel so.
> Also make sure log_warnings=2 which will give more error messages in error
> log.
>
> This is the basic step for this error, if it is not working then we should
> check for other options then.
>
> Regards,
> Vinodh.k
>
>
>
> On Fri, May 27, 2011 at 11:05 AM, Claudio Nanniwrote:
>
>> 'MySQL server has gone away'
>> Can be a network problem,
>> Just to increase complexity :)
>> On May 26, 2011 11:03 PM, "Prabhat Kumar" wrote:
>>> I had experience with such type of error, It was due lack of resources
>>> available to MySql, max connections exceeds on the server.
>>> you can write a simple script which will grab and store output of 'show
>>> processlist' every min. and later you cna investigate the issue.
>>>
>>>
>>> On Wed, May 25, 2011 at 3:34 AM, Aveek Misra
>> wrote:
>>>> Nothing in the error log or the slow query log that suggests that the
>> query
>>>> size is too large or us taking too much time to execute.
>>>>
>>>> Thanks
>>>> Aveek
>>>>
>>>> On May 25, 2011, at 3:53 PM, Rik Wasmus wrote:
>>>>
>>>>>> failed to execute " SELECT * FROM cluster_info WHERE cluster =
>>>> ?":
>>>>>> MySQL server has gone away
>>>>>>
>>>>>> The error "MySQL server has gone away" is the error from the db
>> handle.
>>>> Can
>>>>>> anyone give me any pointers on why that happens? I looked up the
>>>>>> documentation in MySQL docs and the most common reason seems to be
>> that
>>>> it
>>>>>> happens if the query size is very large or if there is a timeout.
>> None
>>>> of
>>>>>> them seems to be a probable cause. The "max_allowed_packet" on the
>>>> server
>>>>>> is 16 MB and as can be seen in the query above, the query is very
>> small
>>>>>> and nowhere near the size limit. We also have a timeout setting
>>>>>> (wait_timeout) of 10 minutes and the above query for us cannot
>> possibly
>>>>>> take that amount of time. In any case, given the same query, it
>> executes
>>>>>> correctly 99% of time (so to speak). It fails intermittently with the
>>>>>> above error. What possibly could be the reason? I also looked at the
>> max
>>>>>> connections on the server at that time (around ~80) and it is much
>> less
>>>>>> than the limit we have (limit is 1000). How can I extract more
>>>> information
>>>>>> when this happens? This error message sucks since it does not tell me
>>>> what
>>>>>> exactly happened. The server version is 5.1.45.
>>>>> Can you access the error log of the server? That can probably shed
>> more
>>>> light
>>>>> on the issue...
>>>>> --
>>>>> Rik Wasmus
>>>>>
>>>>> --
>>>>> MySQL General Mailing List
>>>>> For list archives: http://lists.mysql.com/mysql
>>>>> To unsubscribe:
>>>> http://lists.mysql.com/mysql?unsub=aveekm@yahoo-inc.com
>>>>
>>>> --
>>>> MySQL General Mailing List
>>>> For list archives: http://lists.mysql.com/mysql
>>>> To unsubscribe:
>>>> http://lists.mysql.com/mysql?unsub=aim.prabhat@gmail.com
>>>>
>>>>
>>>
>>> --
>>> Best Regards,
>>>
>>> Prabhat Kumar
>>> MySQL DBA
>>>
>>> My Blog: http://adminlinux.blogspot.com
>>> My LinkedIn: http://www.linkedin.com/in/profileprabhat


--
MySQL General Mailing List
For list archives: http://lists.mysql.com/mysql
To unsubscribe: http://lists.mysql.com/mysql?unsub=gcdmg-mysql-2@m.gmane.org

Re: MySQL server has gone away

Are you using connection pooling in your application?

Regards,
Chandru

On 5/27/2011 11:05 AM, Claudio Nanni wrote:
> 'MySQL server has gone away'
> Can be a network problem,
> Just to increase complexity :)
> On May 26, 2011 11:03 PM, "Prabhat Kumar" wrote:
>> I had experience with such type of error, It was due lack of resources
>> available to MySql, max connections exceeds on the server.
>> you can write a simple script which will grab and store output of 'show
>> processlist' every min. and later you cna investigate the issue.
>>
>>
>> On Wed, May 25, 2011 at 3:34 AM, Aveek Misra wrote:
>>
>>> Nothing in the error log or the slow query log that suggests that the
> query
>>> size is too large or us taking too much time to execute.
>>>
>>> Thanks
>>> Aveek
>>>
>>> On May 25, 2011, at 3:53 PM, Rik Wasmus wrote:
>>>
>>>>> failed to execute " SELECT * FROM cluster_info WHERE cluster =
>>> ?":
>>>>> MySQL server has gone away
>>>>>
>>>>> The error "MySQL server has gone away" is the error from the db
> handle.
>>> Can
>>>>> anyone give me any pointers on why that happens? I looked up the
>>>>> documentation in MySQL docs and the most common reason seems to be
> that
>>> it
>>>>> happens if the query size is very large or if there is a timeout. None
>>> of
>>>>> them seems to be a probable cause. The "max_allowed_packet" on the
>>> server
>>>>> is 16 MB and as can be seen in the query above, the query is very
> small
>>>>> and nowhere near the size limit. We also have a timeout setting
>>>>> (wait_timeout) of 10 minutes and the above query for us cannot
> possibly
>>>>> take that amount of time. In any case, given the same query, it
> executes
>>>>> correctly 99% of time (so to speak). It fails intermittently with the
>>>>> above error. What possibly could be the reason? I also looked at the
> max
>>>>> connections on the server at that time (around ~80) and it is much
> less
>>>>> than the limit we have (limit is 1000). How can I extract more
>>> information
>>>>> when this happens? This error message sucks since it does not tell me
>>> what
>>>>> exactly happened. The server version is 5.1.45.
>>>> Can you access the error log of the server? That can probably shed more
>>> light
>>>> on the issue...
>>>> --
>>>> Rik Wasmus
>>>>
>>>> --
>>>> MySQL General Mailing List
>>>> For list archives: http://lists.mysql.com/mysql
>>>> To unsubscribe:
>>> http://lists.mysql.com/mysql?unsub=aveekm@yahoo-inc.com
>>>
>>> --
>>> MySQL General Mailing List
>>> For list archives: http://lists.mysql.com/mysql
>>> To unsubscribe:
>>> http://lists.mysql.com/mysql?unsub=aim.prabhat@gmail.com
>>>
>>>
>>
>> --
>> Best Regards,
>>
>> Prabhat Kumar
>> MySQL DBA
>>
>> My Blog: http://adminlinux.blogspot.com
>> My LinkedIn: http://www.linkedin.com/in/profileprabhat


--
MySQL General Mailing List
For list archives: http://lists.mysql.com/mysql
To unsubscribe: http://lists.mysql.com/mysql?unsub=gcdmg-mysql-2@m.gmane.org