Re: [Gluster-users] EC clarification
> 2016-09-21 20:56 GMT+02:00 Serkan Çoban : > > Then you can use 8+3 with 11 servers. > > Stripe size won't be good: 512*(8-3) = 2560 and not 2048 (or multiple) It's not really 512*(8+3) though. Even though there are 11 fragments, they only contain 8 fragments' worth of data. They just encode it with enough redundancy that *any* 8 contains the whole. ___ Gluster-users mailing list Gluster-users@gluster.org http://www.gluster.org/mailman/listinfo/gluster-users
Re: [Gluster-users] EC clarification
2016-09-21 20:56 GMT+02:00 Serkan Çoban : > Then you can use 8+3 with 11 servers. Stripe size won't be good: 512*(8-3) = 2560 and not 2048 (or multiple) ___ Gluster-users mailing list Gluster-users@gluster.org http://www.gluster.org/mailman/listinfo/gluster-users
Re: [Gluster-users] EC clarification
Then you can use 8+3 with 11 servers. On Wed, Sep 21, 2016 at 9:17 PM, Gandalf Corvotempesta wrote: > 2016-09-21 16:13 GMT+02:00 Serkan Çoban : >> 8+2 is recommended for 10 servers. From n+k servers it will be good to >> choose n with a power of 2(4,8,16,vs) >> You need to add 10 bricks if you want to extend the volume. > > 8+2 means at least 2 failed bricks, right? > That's too low. I need at least 3 bricks ___ Gluster-users mailing list Gluster-users@gluster.org http://www.gluster.org/mailman/listinfo/gluster-users
Re: [Gluster-users] EC clarification
2016-09-21 16:13 GMT+02:00 Serkan Çoban : > 8+2 is recommended for 10 servers. From n+k servers it will be good to > choose n with a power of 2(4,8,16,vs) > You need to add 10 bricks if you want to extend the volume. 8+2 means at least 2 failed bricks, right? That's too low. I need at least 3 bricks ___ Gluster-users mailing list Gluster-users@gluster.org http://www.gluster.org/mailman/listinfo/gluster-users
Re: [Gluster-users] EC clarification
8+2 is recommended for 10 servers. From n+k servers it will be good to choose n with a power of 2(4,8,16,vs) You need to add 10 bricks if you want to extend the volume. On Wed, Sep 21, 2016 at 4:12 PM, Gandalf Corvotempesta wrote: > 2016-09-21 14:42 GMT+02:00 Xavier Hernandez : >> You *must* ensure that *all* bricks forming a single disperse set are placed >> in a different server. There are no 4 special fragments. All fragments have >> the same importance. The way to do that is ordering them when the volume is >> created: >> >> gluster volume create test disperse 16 redundancy 4 >> server{1..20}:/bricks/test1 server{1..20}:/bricks/test2 >> server{1..20}:/bricks/test3 >> >> This way all 20 fragments from each disperse set will be placed in a >> different server. However each server will have 3 bricks and no fragment >> from a single file will be stored in more than one brick of each server. > > Now it's clear. > So, at very minimum, EC is good starting from 7+3 (10 servers with 1 > brick each) because: 512*(7-3) = 2048 > Any smaller combinations would mean less redundancy (6+2) or not > optimized stripe size like 512*(5-3)=1024 > > Is this correct ? And what if I have to add some brick to the current > servers or I have to add new servers? > Can I add them freely or I have to follow some rules ? > ___ > Gluster-users mailing list > Gluster-users@gluster.org > http://www.gluster.org/mailman/listinfo/gluster-users ___ Gluster-users mailing list Gluster-users@gluster.org http://www.gluster.org/mailman/listinfo/gluster-users
Re: [Gluster-users] EC clarification
2016-09-21 14:42 GMT+02:00 Xavier Hernandez : > You *must* ensure that *all* bricks forming a single disperse set are placed > in a different server. There are no 4 special fragments. All fragments have > the same importance. The way to do that is ordering them when the volume is > created: > > gluster volume create test disperse 16 redundancy 4 > server{1..20}:/bricks/test1 server{1..20}:/bricks/test2 > server{1..20}:/bricks/test3 > > This way all 20 fragments from each disperse set will be placed in a > different server. However each server will have 3 bricks and no fragment > from a single file will be stored in more than one brick of each server. Now it's clear. So, at very minimum, EC is good starting from 7+3 (10 servers with 1 brick each) because: 512*(7-3) = 2048 Any smaller combinations would mean less redundancy (6+2) or not optimized stripe size like 512*(5-3)=1024 Is this correct ? And what if I have to add some brick to the current servers or I have to add new servers? Can I add them freely or I have to follow some rules ? ___ Gluster-users mailing list Gluster-users@gluster.org http://www.gluster.org/mailman/listinfo/gluster-users
Re: [Gluster-users] EC clarification
On 21/09/16 14:36, Gandalf Corvotempesta wrote: Il 01 set 2016 10:18 AM, "Xavier Hernandez" mailto:xhernan...@datalab.es>> ha scritto: If you put more than one fragment into the same server, you will lose all the fragments if the server goes down. If there are more than 4 fragments on that server, the file will be unrecoverable until the server is brought up again. Putting more than one fragment into a single server only makes sense to account for disk failures, since the protection against server failures is lower. Exactly, what i would like to ensure is that the 4 segments needed for recovery are placed automatically on at least 4 servers (and not on 4 different bricks that could be on the same server) You *must* ensure that *all* bricks forming a single disperse set are placed in a different server. There are no 4 special fragments. All fragments have the same importance. The way to do that is ordering them when the volume is created: gluster volume create test disperse 16 redundancy 4 server{1..20}:/bricks/test1 server{1..20}:/bricks/test2 server{1..20}:/bricks/test3 This way all 20 fragments from each disperse set will be placed in a different server. However each server will have 3 bricks and no fragment from a single file will be stored in more than one brick of each server. Xavi ___ Gluster-users mailing list Gluster-users@gluster.org http://www.gluster.org/mailman/listinfo/gluster-users
Re: [Gluster-users] EC clarification
Il 01 set 2016 10:18 AM, "Xavier Hernandez" ha scritto: > If you put more than one fragment into the same server, you will lose all the fragments if the server goes down. If there are more than 4 fragments on that server, the file will be unrecoverable until the server is brought up again. > > Putting more than one fragment into a single server only makes sense to account for disk failures, since the protection against server failures is lower. > Exactly, what i would like to ensure is that the 4 segments needed for recovery are placed automatically on at least 4 servers (and not on 4 different bricks that could be on the same server) ___ Gluster-users mailing list Gluster-users@gluster.org http://www.gluster.org/mailman/listinfo/gluster-users
Re: [Gluster-users] EC clarification
Hi, On 27/08/16 10:57, Gandalf Corvotempesta wrote: In short: how can i set the node hosting the erasure codes? In a 16+4 EC (or bigger) i would like to put the 4 bricks hosting the ECs on 4 different servers so that i can loose 4 servers and still be able to access/recover data EC builds several fragments of data for each file. In the case of a 16+4 configuration, a file of size 1MB is transformed into 20 smaller files (fragments) of 64KB. To recover the original file *any* subset of 16 fragments is enough. There aren't special fragments with more information or importance. If you put more than one fragment into the same server, you will lose all the fragments if the server goes down. If there are more than 4 fragments on that server, the file will be unrecoverable until the server is brought up again. Putting more than one fragment into a single server only makes sense to account for disk failures, since the protection against server failures is lower. Xavi ___ Gluster-users mailing list Gluster-users@gluster.org http://www.gluster.org/mailman/listinfo/gluster-users ___ Gluster-users mailing list Gluster-users@gluster.org http://www.gluster.org/mailman/listinfo/gluster-users
[Gluster-users] EC clarification
In short: how can i set the node hosting the erasure codes? In a 16+4 EC (or bigger) i would like to put the 4 bricks hosting the ECs on 4 different servers so that i can loose 4 servers and still be able to access/recover data ___ Gluster-users mailing list Gluster-users@gluster.org http://www.gluster.org/mailman/listinfo/gluster-users