[Gluster-users] about split-brain

[Roman]

Personally I would call it SPOF (single point of failure) and you should
take care of it, when you work with some HA software.
HA requires not only software, but hardware HA (double switches, power
sources etc). But, maybe you will have some help from devs for this.


2014-08-08 9:44 GMT+03:00 hwh at sangfor.com <hwh at sangfor.com>:

> because if the network is not stable, this will be ocurr and file can't
> open due to split brain
>
> ------------------------------
> hwh at sangfor.com
>
>
> *From:* Roman <romeo.r at gmail.com>
> *Date:* 2014-08-08 14:27
> *To:* hwh at sangfor.com
> *Subject:* Re: Re: [Gluster-users] about split-brain
> How can it be a bug, if you don't wait for synchronization to complete?
>
>
> 2014-08-08 9:24 GMT+03:00 hwh at sangfor.com <hwh at sangfor.com>:
>
>> Is glusterfs bug ?  i want to know whether have some way to avoid this
>> split-brain?
>>
>> ------------------------------
>> hwh at sangfor.com
>>
>>
>> *From:* Roman <romeo.r at gmail.com>
>> *Date:* 2014-08-08 14:16
>> *To:* hwh at sangfor.com
>> *CC:* gluster-users <gluster-users at gluster.org>
>> *Subject:* Re: Re: [Gluster-users] about split-brain
>> So try your tests again, but wait for synchronization first. You can
>> check the state of files on servers with getfattr -d -m. -e hex
>> /path-to-file-on-gluster-server
>> both should return 0x000000000000000000000000
>>
>>
>> 2014-08-08 9:11 GMT+03:00 hwh at sangfor.com <hwh at sangfor.com>:
>>
>>> Yes, I do not wait for synchronization to complete.
>>>
>>> ------------------------------
>>> hwh at sangfor.com
>>>
>>>
>>> *From:* Roman <romeo.r at gmail.com>
>>> *Date:* 2014-08-08 13:53
>>> *To:* hwh at sangfor.com
>>> *CC:* gluster-users <gluster-users at gluster.org>
>>> *Subject:* Re: [Gluster-users] about split-brain
>>> Hello,
>>>
>>> Did you wait for synchronization to complete before plugging second
>>> server off? Usually it need 10-15 minutes. Sometimes faster.
>>>
>>>
>>> 2014-08-08 5:19 GMT+03:00 hwh at sangfor.com <hwh at sangfor.com>:
>>>
>>>>  hi:
>>>>
>>>> I have a problem about split-brain, details as below:
>>>>
>>>> PC1 and PC2 is AFR, stored a file 1.txt
>>>>
>>>> 1. write "123" to 1.txt, then 1.txt in pc1 is same with 1.txt in PC2,
>>>> content is "123"
>>>>
>>>> 2. PC1  network disconnect, write "45" to 1.txt, then  content of 1.txt
>>>> in pc1 is "123",
>>>>
>>>>    content of 1.txt in pc2 is "12345"
>>>>
>>>> 3. PC2  network disconnect, but PC1 network is good, write "67" to
>>>> 1.txt, then  content of 1.txt in pc1 is "12367",
>>>>
>>>>    content of 1.txt in pc1 is "12345"
>>>>
>>>> 4. so split brain occur
>>>>
>>>>
>>>> how glusterfs avoid this? Have some ways to solve? this is my
>>>> questions. thank you
>>>>
>>>> ------------------------------
>>>> hwh at sangfor.com
>>>>
>>>> _______________________________________________
>>>> Gluster-users mailing list
>>>> Gluster-users at gluster.org
>>>> http://supercolony.gluster.org/mailman/listinfo/gluster-users
>>>>
>>>
>>>
>>>
>>> --
>>> Best regards,
>>> Roman.
>>>
>>>
>>
>>
>> --
>> Best regards,
>> Roman.
>>
>>
>
>
> --
> Best regards,
> Roman.
>
>


-- 
Best regards,
Roman.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://supercolony.gluster.org/pipermail/gluster-users/attachments/20140808/f40c9fb6/attachment.html>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: CatchC5D8(08-08-(08-08-14-41-48).jpg
Type: image/jpeg
Size: 13852 bytes
Desc: not available
URL: <http://supercolony.gluster.org/pipermail/gluster-users/attachments/20140808/f40c9fb6/attachment.jpg>