[DRBD-user] DRBD write after write (cache)

Felix Frank ff at mpexnet.de
Wed Feb 9 12:46:22 CET 2011

Note: "permalinks" may not be as permanent as we would like,
direct links of old sources may well be a few messages off.


On 02/09/2011 12:21 PM, ionral wrote:
> 
> 
> 
> Felix Frank-2 wrote:
>>
>> Hi.
>>
>>> if I check the status of drbd the following response is: 
>>>
>>>  0: cs: Connected ro: Primary / Primary ds: UpToDate / UpToDate C r ---- 
>>>     ns: 640864680 nr: 135234784 dw: 776099464 dr: 1520599328 al: 836478
>>> bm:
>>> 2185 lo: 0 pe: 0 ua: 0 ap: 0 ep: 1 wo: b OOS: 0 
>>>
>>> I do not know if it is positive that the method is to write after is
>>> Barrier
>>> .... reading the manuals of drbd I noticed that in these cases (BBU)
>>> performance would be better for a wo:n 
>>>
>>> What do you think? 
>>
>>>> Possibly? You may want to just test the performance difference.
>>
>> so there is no certainty that it is more powerful
>> I can just do a test


I didn't say that. *I* don't have any certainty because I never tried
it. If the documentation states that there is a performance gain, there
is no reason to believe otherwise. I'm just not handing out
confirmations about things I'm clueless about.

>>
>>> more to this I can tell you that another hardware configuration with a
>>> simple SATA disks in software raid  0 configuration (sda and sdb) and
>>> without drbd.conf options: no-disk-flushes -no-md-flushes, The drbd uses
>>> instead the drain ... that seems to be dangerous without a BBU, but the
>>> drbd
>>> has preferred this choice, and not clear to me ...
>>
>>>> Check you kernel logs. DRBD may have noticed that barriers are not
>>>> available for your backing device and thus falls back to drain. Is LVM
>>>> involved in your that setup per chance?
>>
>>>> Regards,
>>>> Felix
>>
>> Unfortunately I have nothing in the logs, the rotate deleted the old log,
>> but do not understand much because the drbd not find in my backing storage 
>> support for the barrier, I use LVM on top of drbd and not vice versa, is
>> not this strange thing?

The setup is fine. If it's not painful for you, you may want to bring
the resource down and up again, watch the /proc/drbd output and if it
drops from using flush, check your logs again. Except someone comes up
with a better idea about why you're on drain.

HTH,
Felix



More information about the drbd-user mailing list