[ome-devel] Trouble with OMERO.Dropbox
Colin Blackburn
c.blackburn at dundee.ac.uk
Sat Feb 14 12:05:05 GMT 2015
Hi Niko,
> I did a little research by myself, alas, I couldn't find much in this direction. So I'd definitely be interested in your findings - especially as I'm using inotify myself in another project.
There are some options that define how pyinotify interacts with inotify that I’m looking at. I’ll try to update you next week and point you at a branch that you can patch from if it is worth testing further.
> I'll try to find some time to figure out a reliable way to reproduce the super-high load on our OMERO when using DropBox. I did have the impression that it's worse with some file formats (e.g. the LIF from the logs) compared to others. But right now this is only a gut feeling, so please don't rely on this.
If you can find the time, then any numbers you can provide would be most welcome.
Cheers,
Colin
> On 12 Feb 2015, at 08:25, Niko Ehrenfeuchter <nikolaus.ehrenfeuchter at unibas.ch> wrote:
>
> Hi Colin,
>
> On 06.02.2015 11:08, Colin Blackburn wrote:
>> Hi Niko,
>>
>>> Hum, I thought more of something at OS level for configuring
>>> inotify itself.
>>
>> Yes, but at the present time any configuration there is, is only
>> available to us through the Python bindings. I will, though, look
>> into this further.
>
> I did a little research by myself, alas, I couldn't find much in this direction. So I'd definitely be interested in your findings - especially as I'm using inotify myself in another project.
>
>
>>> Regarding the Python bindings, are you aware of the following?
>>>
>>> *
>>> http://www.serpentine.com/blog/2008/01/04/why-you-should-not-use-pyinotify/
>>> (particularly interesting regarding performance is also the last
>>> comment) * https://bitbucket.org/bos/python-inotify (original
>>> repo) * https://bitbucket.org/JanKanis/python-inotify/ (more recent
>>> fork)
>>>
>>> There are a couple of additional forks of the latter, whereof the
>>> only one containing significant changes is the "poupas" one.
>>
>> Yes, at the time we adopted pyinotify it was done for pragmatic
>> reasons. We’ll certainly look at whether we can somehow allow for
>> other bindings to the notification system to be used in the future.
>
> Nice! Again, I'm always interested in your insights there.
>
>> More generally, I’m investigating the problem you are seeing on one
>> of our systems here. I’ll get back to you about that as soon as I
>> can.
>
> Great :)
>
> I'll try to find some time to figure out a reliable way to reproduce the super-high load on our OMERO when using DropBox. I did have the impression that it's worse with some file formats (e.g. the LIF from the logs) compared to others. But right now this is only a gut feeling, so please don't rely on this.
>
> Cheers & thanks again!
> Niko
> _______________________________________________
> ome-devel mailing list
> ome-devel at lists.openmicroscopy.org.uk
> http://lists.openmicroscopy.org.uk/mailman/listinfo/ome-devel
The University of Dundee is a registered Scottish Charity, No: SC015096
More information about the ome-devel
mailing list