On 07/14/15 17:47, Jesse Stroik wrote:
> Thomas,
>
> What we'd like to do is set up a reliable check_mk alert. We now rely
> on robinhood for managing certain aspects of our file systems. If it
> is down overnight that's significant.
>
> We'd been monitoring the changelogs but given the batching, it's not
> sufficiently responsive and it feels like a hack which we'd like to
> purge.
>
> An idea we have is to modify the robinhood startup and use that to
> report the status of the file systems.
Hi Jesse,
Can you explain in what fashion you want to modify robinhood startup? I
am not sure to understand.
Thomas
> But before doing that, it seems prudent to ask you if you'd recommend
> a different method because we want to take the best tack.
>
> Best,
> Jesse Stroik
>
>
> On 2/11/2015 3:12 AM, LEIBOVICI Thomas wrote:
>> Hi Scott,
>>
>> The information reported by "rbh-report -a" is not updated in real-time:
>> it is updated at regular interval (specified by "log::stats_interval"
>> config parameter).
>> Moreover, regarding you file creation test, there can be a delay between
>> the creation action and the time robinhood actually process it,
>> due to internal changelog batching mechanism, driven by:
>> changelog::queue_max_size, changelog::queue_max_age,
>> changelog::queue_check_interval, and also the delay to acknowledge
>> changelog records to Lustre driven by changelog::batch_ack_count.
>>
>> To check robinhood correctly reads the changelog, you can monitor the
>> delta between the last record id and the last acknowledged record.
>> You can check this on Lustre MDS :
>>
>> cat /proc/fs/lustre/mdd/*-MDT0000/changelog_users
>> current index: 124950267
>> ID index
>> cl1 124950266
>>
>> Here you can see that current index is very close to the last record
>> acknowledged by cl1.
>>
>> There are other ways to monitor robinhood changelog processing, like
>> defining a dump file to dump all received changelogs
>> ("changelog::dump_file" config parameter), running robinhood with DEBUG
>> level, etc...
>>
>> Regards
>> Thomas
>>
>>
>> On 02/10/15 19:16, Scott Nolin wrote:
>>> Hello,
>>>
>>> We have been trying to monitor robinhood by looking at the 'Last read
>>> record time' in 'rbh-report -a' activity status.
>>>
>>> Assuming our filesystems are sometimes just not doing anything, part
>>> of this check first creates a new file on the filesystem.
>>>
>>> I've found that this often shows no change for hours. Am I missing
>>> something simple here?
>>>
>>> What's the best way to see if robinhood is properly working with your
>>> changelogs?
>>>
>>> Thanks,
>>> Scott
>>>
>>>
------------------------------------------------------------------------------
Don't Limit Your Business. Reach for the Cloud.
GigeNET's Cloud Solutions provide you with the tools and support that
you need to offload your IT needs and focus on growing your business.
Configured For All Businesses. Start Your Cloud Today.
https://www.gigenetcloud.com/
_______________________________________________
robinhood-support mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/robinhood-support
