[raxxeh]

Hope i dont have to deal with this when i finally get my mSP...

You won't.

Just use the one kernel that is stable, forget about using virtualization, and you're golden.

[jonlewi5]

Hope i dont have to deal with this when i finally get my mSP...

[Phixion]

Thank you OVH!

They have changed my motherboard for an Intel with Intel NIC and everything is rock solid so far, I'm so glad to see the back of that network card.

My faith is restored!

[Phixion]

Happy to say they are now changing my motherboard (hoping to god it has a different NIC!).

[JakeMS]

Hi

I can confirm the RAM and Power Supply were replaced I can see you have replied with logs showing that eth0 still keeps failing so we continue to find the source of the problem.

It's the NIC...

The realtek does not play nice with Linux, just stick a couple of Intel PRO/1000 MT's in these servers and use them instead, and all the dropouts etc, will stop.

Simple as that..

[Phixion]

Everyone I know has the same NIC and most of them have this exact problem.

I think you're lucky if you don't have this problem.

[jonlewi5]

Do we know if this is the hardware supplied in all mSP's?

[Phixion]

OK so I have had a stable server overnight with no connection drops or crashes.

It seems the only Kernel I can use is:

bzImage-3.10.11-r8168-grs-ipv6-64

Any other Kernel:

Default = Crash
Custom = Crash
Proxmox = Crash
bzImage-3.10.11-r8168-grs-ipv6-64 = Stable
bzImage-3.10.12-r8168-grs-ipv6-64 = Connection dropping

So as it stands I am stuck with 1 Kernel, forever. Any chance of me using Proxmox or any other Default/Custom Kernel is slim to none.

I don't think I'm out of line to point to the NIC being the source of the problem here, everything but the NIC/Mobo/HDDs have been swapped out.

[Neil]

That is the problem though, there is no other provider that offers the same level of hardware for the price.

I've been with OVH since 2009 and their servers have never failed me... until now.

I am just frustrated that none of the techs will admit that the network card is crap and change it for me.

Almost 2 months of this and my patience is starting to wear thin.

OK support is telling me they have replace RAM/PSU, even though my ticket said otherwise.

Time to test... again

Hi

I can confirm the RAM and Power Supply were replaced I can see you have replied with logs showing that eth0 still keeps failing so we continue to find the source of the problem.

[Kacotet]

Drop the server and order a new one when the order system is open.

[Phixion]

That is the problem though, there is no other provider that offers the same level of hardware for the price.

I've been with OVH since 2009 and their servers have never failed me... until now.

I am just frustrated that none of the techs will admit that the network card is crap and change it for me.

Almost 2 months of this and my patience is starting to wear thin.

OK support is telling me they have replace RAM/PSU, even though my ticket said otherwise.

Time to test... again

[JakeMS]

I know it sucks to say it, but after all that, I would suggest you try to find a provider who can provide a working server which can suit your needs.

[Phixion]

I have to say I am very disappointed in the support I have received from OVH regarding this issue.

My server (mSP) came with an ASUS motherboard and a Realtek NIC. Upon receiving the server I noticed the server would crash at least once a day and I had many interventions over the 1st months rental.

I managed to stumble upon a thread on these forums regarding the issue, there are also various threads around the internet from people with the same issue.

Apparently the Realtek 8168 doesn't play well with certain Linux Kernels, the only "fix" that OVH could offer was to use an OVH 8168 patched Kernel, not really a fix if you plan on using custom Kernels or things like Proxmox.

I had a ticket back and forth with Felix that was closed after I agreed to use the OVH 8168 Kernel. Everything worked fine for a while until I noticed my connection would drop, sometimes it would kick me off SSH and other times SSH would just freeze for 5 seconds, this was evident in the syslog:

cat /var/log/syslog

Sep 30 02:58:17 kernel: r8168: eth0: link down
Sep 30 02:58:21 kernel: r8168: eth0: link up
Sep 30 03:00:15 kernel: r8168: eth0: link down
Sep 30 03:00:19 kernel: r8168: eth0: link up

So, I open another ticket on the 30th September, this time I have to go over everything again and ask them to change the Mobo/NIC to another model because this network card obviously doesn't work well in Linux.

The technician carried out some tests and eventually arranged an intervention:

I suggest in this case to replace your RAM and power supply
and update your bios.

If you agree, please fix the date and time of the intervention.

I responded and told them to carry out the intervention as soon as possible as this was now the 4th day of waiting for this issue to be resolved.

So, the intervention was planned, I receive a ticket:

An operation was added for ********.ovh.net to the list of
interventions at 2013-10-03 15:50:00.

Here are the details of this operation: BIOS check

You will be notified when work will begin shortly before
the intervention.

BIOS check?! What happened to RAM and PSU swap?

The intervention on *******.ovh.net has been completed.

This operation was closed at 2013-10-03 19:35:47

Here are the details of this operation:
BIOS check

Great... a BIOS check. By this time it's already too late to phone OVH technical support and I'm left wondering what to do.

So for 2 months I've had intermittent connection to server, multiple crashes, 4 days after submitting a ticket all I've got out of it is a BIOS check. The RAM and PSU replacement (which I don't think will change anything) has been somehow forgotten about.

I've phoned technical support on numerous occasions only to be told that I have to submit a ticket and wait for a reply, they can't do anything for me. In that case, why does technical support even exist?

After I questioned the ethics of knowingly selling servers with faulty hardware and not following their own SLA, a gentlemen at technical support told me that I was lucky my ticket was even replied to because "you usually have to pay to get a reply for these things" - which blew my mind, I'm pretty sure it says if the fault is not your own you don't have to pay.

To handle an incident, OVH make a diagnosis. If the diagnosis reveals that the incident is the responsibility of OVH, the resolution of the incident is part of your warranty service. Otherwise, the diagnosis, which may be accompanied by a quote, you will be charged £20.

In case of hardware problems, the faulty part is replaced.
We can replace power supply, a hard disk, RAM or the whole server is substituted.

The SLA agreement speaks volumes to me, OVH obviously can't or won't replace my server with a new one because it involves swapping the entire server, instead they promise to replace parts and fail to do so.

This is not good enough OVH.