Posts by omvault

    Lol, clearly the personas are not yet landed and since contextually I guess I simply misinterpreted the remarks since Agricola said some stuff first re: the point in question.


    ryecoaaron I interpret your remark as "I just *DON'T* have a good way to quantify that" vs the positive rendition, right? ;)


    Definitely would be interesting to see some actual metrics for both. I agree ryecoaaron, they are different and one would be measured by "inside" benchmarking while the other would be measured from outside (Proxmox) process indicators (htop for a quickie or other more memory process-memory measurements during process lifetime).


    ness1602 yeah, thanks, I too have the same kind of impression / thoughts. I really think this topic is interesting (and important) since everything so far everywhere just seems to be impressions and beliefs vs specific metrics. <shrug>

    ryecoaaron who's Volker? I'm inferring that it's Agricola in context, but not clear.


    As to specific performance concerns, I have perhaps an identical background that you have as per your brief history synopsis. I simply always look for ways to improve performance / reduce processor utilization + energy requirements (via pushing more services to a system via some form of virtualization). When I can squeeze more (sane) processes with less utilization onto a system I do it and thus my interest.


    So far I've been shooting blanks of finding solid performance comparisons between KVM and LXC, but I'll post it here if/when I do happen across something that I feel is pertinent.

    Anyone familiar with any updated or more recent guides that cover the steps and any suggestions that are covered in guides such as these?:

    Working on setting up an Unraid-like solution that offers the benefits of flexible expansion, but incorporates the bit-rot protection and recovery of SnapRAID, etc.

    ryecoaaron I think in this age of ever increasing software bloat and lack of optimization considerations it is important to take a step back and try to obtain the most and best performance with lowest resource usage possible. I know many share this same belief and I always try to make such decisions vs the other way of thinking.


    Can you tell me the performance impact of a vm guest approach vs a containerized approach with LXC, for example? I would really love to see some benchmarks/measurements to just differentiate them. I would really quite like to see that the differences are as small as you believe they are.


    Frankly I'm running in a VM on a pretty fat server and passthrough PCIs (HBA and USB). I want/think the isolation and clarity + restorability are probably worthwhile, BUT I am still interested in the LXC approach and want to see metrics. I think it is important and not to be discounted/discarded.