Hi Justin,

The HTTPS is not very stable and better if you'll use the HTTP. To change a protocol you should use NMC setup appliance init command.

Hmm, that's unusual, I would have thought that SSL connections would be the recommended method to connect, but OK, I'll keep that in mind.

For the record (or your KB if you so like), I think I have fixed the access response time I mentioned.

Recently, I upgraded the system with a new mainboard and CPU. This change also included a new network interface. When I was looking in the syslogs today, I noted that the system was complaining that the old interface name no longer existed. Checking the network setup, the new interface name was not set as the primary interface. Correcting this seems to have cleared up the hesitation.

Thank you for the followup

it works much metter with ie7 for whatever reason ...

And that is sad, IMO.

I try to avoid software from Redmond, and primary run Linux or Mac OS, so using IE isn't something that I can do easily, if at all.

We will improve HTTPS and make it 100% stable in 3.0.4. I think 3.0.4 be available in July, now we're working on 3.0.3 and our main aim is stable kernel and perfomance.

That is good to hear, thank you for the update Pavel.

I do have to give you guys/gals credit and good word of mouth, as it were. Many of us, I'm sure, who are using these forums are using the free community edition or the time limited demo, and when issues have appeared, your company has been fairly responsive (better than some companies I've had to deal with in the past, even with the 'Gold 2000 Pro' support package).

I look forward to the continuing improvements.

Freshly installed 3.0.3, first http page 'wizard1' doesn't load, timesout on firefox36, same on ie7. Too bad this project's not working yet, I'll check back in another 6 months.

Wow, nice post. No real troubleshooting listed, just 'I tried this and it doesn't work, the whole thing is broken'.

shakes head

Tried again using rtl8139 instead of e1000 which is what was used during the fail, interface came up immediateley.

So much for jumbo frames. Only e1000 or virtio can do >1500 in qemu/KVM, and NexentaStor supports neither AFAIK.

Even if I give it direct access to disk, it's still less optimal than I'd like for it to be. Seems if it's not one thing it's another.

Maybe I can just team a few & acheive the same, but then there's the overhead associated with a bunch of fully virtualized interfaces. I wonder if that's much, or if it'd even help throughput, hmm.