ignite-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Petr Ivanov <mr.wei...@gmail.com>
Subject Re: Upsource lags and connection timeouts
Date Tue, 09 Oct 2018 14:43:07 GMT
New dedicated disk subsystem is installed for MTCGA utility, hope that'll help to free resources
for other services.


> On 21 Sep 2018, at 21:10, Anton Vinogradov <av@apache.org> wrote:
> 
> Folks,
> 
> Thanks for the investigation!
> Could you please let us know once the problem will be solved?
> 
> пт, 21 сент. 2018 г. в 19:28, Dmitriy Pavlov <dpavlov.spb@gmail.com>:
> 
>> Hi Anton, Petr,
>> 
>> I confirm that TC Bot writes a lot of stuff, so it may create problems for
>> Cassandra storage.
>> 
>> The latest version of Bot has optimization of data saving, and its reduces
>> amount of data to be saved to disk. But still, disc separation is a better
>> solution.
>> 
>> Sincerely,
>> Dmitriy Pavlov
>> 
>> пт, 21 сент. 2018 г. в 19:19, Petr Ivanov <mr.weider@gmail.com>:
>> 
>>> Anton,
>>> 
>>> 
>>> We shared this issue with Jetbrains and they found no current problems in
>>> logs, except possible problems with database connection (Apache
>> Cassandra).
>>> Currently there are no errors in logs.
>>> 
>>> Possible troublemakers though exist:
>>> - MTCGA utility “lives” on the same server, possible resource race
>>> condition, will try to separate services (at least on disk subsystem
>> level)
>>> - connection to GitHub is an issue too — sometimes fetch speed drops
>>> significantly, that will be investigated also
>>> 
>>> 
>>>> On 20 Sep 2018, at 22:18, Anton Vinogradov <av@apache.org> wrote:
>>>> 
>>>> Still, see the same issues.
>>>> 
>>>> Denis,
>>>> 
>>>> Could you please prioritize the fix?
>>>> 
>>>> вт, 18 сент. 2018 г. в 14:01, Anton Vinogradov <av@apache.org>:
>>>> 
>>>>> Also, I see following on each resolve attempt.
>>>>> UpsourceRequestExceptionImpl: 108: Internal error: An error occurred
>>>>> during flushing data to database
>>>>> 
>>>>> 
>>>>> 
>>>>> вт, 18 сент. 2018 г. в 13:56, Anton Vinogradov <av@apache.org>:
>>>>> 
>>>>>> Folks,
>>>>>> 
>>>>>> Who is responsible for Upsouce [1]?
>>>>>> I see a performance issues last week.
>>>>>> Can we check Upsource health?
>>>>>> 
>>>>>> [1] https://reviews.ignite.apache.org
>>>>>> 
>>>>> 
>>> 
>>> 
>> 


Mime
View raw message