Re: [Moosefs-users] Unavailable Chunks

[Stéphane B. <ste...@ga...>]

Its starting to really concering me. Every loop I get missing chunks and 
under-goal files... We have production web server running on this and we 
cannot afford to have unavailable files. When a file is unavailable this 
means it's a lost sell. The numbers are getting higher. Maybe the load 
is too high ?

We are having about 6 server with about 1 millions lookup per hours each 
, 500k opens per hours and 70k read per hours.

We also have another mooseFS cluster with the same settings on different 
machines and we have no under-goal files or missing files.


Thanks



check loop start time 	check loop end time 	files 	under-goal files 
missing files 	chunks 	under-goal chunks 	missing chunks
Thu Mar 17 11:23:52 2011 	Thu Mar 17 15:24:58 2011 	2125751 	25891 
2075 	2119886 	25891 	2075





On 11-03-15 10:24 AM, Stéphane Boisvert wrote:
> Thanks for the answer,
>
> But it happens really often. Maybe every 2 or 3 loops. Maybe it is 
> some settings wrong ? We have 2 chunkserver ... that means that both 
> of them was unavailable. I changed the timeout settings a little.
>
> Here is my settings.
>
>
> Chunkservers:
>
> MASTER_TIMEOUT = 2
> HDD_TEST_FREQ = 10
>
> This is the only 2 settings I changed
>
> Masters:
>
> CHUNKS_LOOP_TIME = 60
> CHUNKS_DEL_LIMIT = 5000
> CHUNKS_WRITE_REP_LIMIT = 5
> CHUNKS_READ_REP_LIMIT = 10
>
> This the only settings I changed the remaining should be Default
>
>
> More explanation about the setup...
>
> We have 2 master running Keepalived (same as carp)
> the chunk servers are on the same servers as the master
> the machines are Dual quad core Xeons with 8G of Ram
> and 15k rpm disks in raid 5
>
>
>
> On 11-03-15 08:50 AM, Michal Borychowski wrote:
>> Hi!
>>
>> These are logs from a test loop. It may happen that while the loop is
>> running some chunkservers are unavailable. The next loop should show that
>> everything is allright.
>>
>> Unless you have numbers in red in the first column in CGI monitor,
>> everything is fine.
>>
>>
>> Kind regards
>> Michał Borychowski
>> MooseFS Support Manager
>> _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
>> Gemius S.A.
>> ul. Wołoska 7, 02-672 Warszawa
>> Budynek MARS, klatka D
>> Tel.: +4822 874-41-00
>> Fax : +4822 874-41-01
>>
>>
>>
>> -----Original Message-----
>> From: Stéphane Boisvert [mailto:ste...@ga...]
>> Sent: Thursday, March 10, 2011 8:40 PM
>> To: moosefs-users
>> Subject: [Moosefs-users] Unavailable Chunks
>>
>> Hi,
>>       We have moosefs running 2 months now. And today we got few errors
>> on it. The master has done the: structure check loop and it ended with
>> 72 unavailable chunks and files.  I don't know what this really means.
>> The files are still accessible  and mfscheckfile output is good too. We
>> have about 2 millions files on the mooseFS currently and 2 chunkserver
>> with a redundant setup. On this cluster we run the version: 1.6.17 we
>> plan to upgrade it soon.
>>
>>
>> Thanks,
>>
>> Stephane
>>
>>
>>
>>
>> ----------------------------------------------------------------------------
>> --
>> Colocation vs. Managed Hosting
>> A question and answer guide to determining the best fit
>> for your organization - today and in the future.
>> http://p.sf.net/sfu/internap-sfd2d
>> _______________________________________________
>> moosefs-users mailing list
>> moo...@li...
>> https://lists.sourceforge.net/lists/listinfo/moosefs-users
>>
>
>
> -- 
>
> 	
>
> *Stephane Boisvert*
> Unix Administrator
>
> Msn : ste...@ga...
> E-mail :ste...@ga...
>
>
>
>
> ------------------------------------------------------------------------------
> Colocation vs. Managed Hosting
> A question and answer guide to determining the best fit
> for your organization - today and in the future.
> http://p.sf.net/sfu/internap-sfd2d
>
>
> _______________________________________________
> moosefs-users mailing list
> moo...@li...
> https://lists.sourceforge.net/lists/listinfo/moosefs-users


-- 

	

*Stephane Boisvert*
Unix Administrator

Msn : ste...@ga...
E-mail :ste...@ga...