I have been using GFS2 with shared mount point in production KVM since long, Trust me you need to have an expert to manage your whole cluster otherwise it becomes very hard to manage, NFS works pretty fine with KVM, if you are planning to use ISCSi or FC, XenServer/XCP and VMware works far far better then KVM and very easy to manage.
Vivek Kumar Sr. Manager - Cloud & DevOps IndiQus Technologies M +91 7503460090 www.indiqus.com > On 29-Oct-2021, at 1:14 PM, Hean Seng <heans...@gmail.com> wrote: > > For primitive way for NFS HA, you can consider is just using DRDB . > > I think is not yet supported linstor here. > > > > On Fri, Oct 29, 2021 at 2:29 PM Piotr Pisz <pi...@piszki.pl> wrote: > >> Hi >> >> So we plan to use linstor in parallel to ceph as a fast resource on nvme >> cards. >> Its advantage is that it natively supports zfs with deduplication and >> compression :-) >> The test results were more than passable. >> >> Regards, >> Piotr >> >> >> -----Original Message----- >> From: Mauro Ferraro - G2K Hosting <mferr...@g2khosting.com> >> Sent: Thursday, October 28, 2021 2:02 PM >> To: users@cloudstack.apache.org; Pratik Chandrakar < >> chandrakarpra...@gmail.com> >> Subject: Re: Experience with clustered/shared filesystems based on SAN >> storage on KVM? >> >> Hi, >> >> We are trying to make a lab with ACS 4.16 and Linstor. As soon as we >> finish the tests we can give you some approach for the results. Are someone >> already try this technology?. >> >> Regards, >> >> El 28/10/2021 a las 02:34, Pratik Chandrakar escribió: >>> Since NFS alone doesn't offer HA. What do you recommend for HA NFS? >>> >>> On Thu, Oct 28, 2021 at 7:37 AM Hean Seng <heans...@gmail.com> wrote: >>> >>>> I have similar consideration when start exploring Cloudstack , but >>>> in reality Clustered Filesystem is not easy to maintain. You seems >>>> have choice of OCFS or GFS2 , gfs2 is hard to maintain and in redhat >>>> , ocfs recently only maintained in oracle linux. I believe you do not >> want to >>>> choose solution that is very propriety . Thus just SAN or ISCSI o is >> not >>>> really a direct solution here , except you want to encapsulate it in >>>> NFS and facing Cloudstack Storage. >>>> >>>> It work good on CEPH and NFS , but performance wise, NFS is better . >>>> And all documentation and features you saw in Cloudstack , it work >>>> perfectly on NFS. >>>> >>>> If you choose CEPH, may be you have to compensate with some >>>> performance degradation, >>>> >>>> >>>> >>>> On Thu, Oct 28, 2021 at 12:44 AM Leandro Mendes >>>> <theflock...@gmail.com> >>>> wrote: >>>> >>>>> I've been using Ceph in prod for volumes for some time. Note that >>>> although >>>>> I had several cloudstack installations, this one runs on top of >>>>> Cinder, but it basic translates as libvirt and rados. >>>>> >>>>> It is totally stable and performance IMHO is enough for virtualized >>>>> services. >>>>> >>>>> IO might suffer some penalization due the data replication inside Ceph. >>>>> Elasticsearch for instance, the degradation would be a bit worse as >>>>> there is replication also in the application size, but IMHO, unless >>>>> you need extreme low latency it would be ok. >>>>> >>>>> >>>>> Best, >>>>> >>>>> Leandro. >>>>> >>>>> On Thu, Oct 21, 2021, 11:20 AM Brussk, Michael < >>>> michael.bru...@nttdata.com >>>>> wrote: >>>>> >>>>>> Hello community, >>>>>> >>>>>> today I need your experience and knowhow about clustered/shared >>>>>> filesystems based on SAN storage to be used with KVM. >>>>>> We need to consider about a clustered/shared filesystem based on >>>>>> SAN storage (no NFS or iSCSI), but do not have any knowhow or >>>>>> experience >>>> with >>>>>> this. >>>>>> Those I would like to ask if there any productive used environments >>>>>> out there based on SAN storage on KVM? >>>>>> If so, which clustered/shared filesystem you are using and how is >>>>>> your experience with that (stability, reliability, maintainability, >>>>> performance, >>>>>> useability,...)? >>>>>> Furthermore, if you had already to consider in the past between SAN >>>>>> storage or CEPH, I would also like to participate on your >>>> considerations >>>>>> and results :) >>>>>> >>>>>> Regards, >>>>>> Michael >>>>>> >>>> >>>> -- >>>> Regards, >>>> Hean Seng >>>> >>> >> >> > > -- > Regards, > Hean Seng
