OVH Community, your new community space.

What to do if there is an emergency?


DigitalDaz
22-12-2014, 15:31
I agree with both Trapper and john_gb here. The best way to approach OVH, and this ranges from Kimsufi right up to higher end OVH models is to assume there is zero support. If you have a problem and it is quickly resolved, you got lucky.

You won't find that many old timers on here I doubt that still use OVH services in a way other than this. I still have a few boxes but everything that is here is part of my playground or can be revived in a UK datacenter very quickly.

The hardware in general is excellent value for money, no one can argue that and you will still be hard pushed to find a good price match anywhere else. OVH have a pretty awesome network too.

What has always let OVH down since the very day I joined them is support. The bad thing is though, most companies that start off with bad support get better. Anyone who has been with OVH for any length of time will tell you its the exact opposite with OVH, its actually got much, much worse over the last five years or so.

Credit where credits due though, I can only remember having had a few hard drive failures and one hardware RAID card go in the last 7 years.

The hardware RAID card if I remember correctly was dodgy from the moment it was provisioned and the server was unusable but it still took me over 36 hours to persuade them to change it.

As for that 24 hour emergency line, it took just about an hour to get an answer on and that is no exaggeration. Once I got through, because it wasn't tripping their hardware alert thing, ie it wasn't completely broken and was still responding to ping, they would not treat it as a level one fault and simply submitted it as a level two which meant I had to wait until later in the day for someone to look at it.

All that happened back in the day when the level of support was reasonable. I think that was on one of the MG servers that was costing me about £200 a month at that time. The server never got renewed after that.

Trapper
19-12-2014, 18:40
Quote Originally Posted by john_gb
Thanks for the support Trapper, but with respect, you have missed the essence of my post, which is, with the right planning and design you CAN have automatic failover and disaster recovery built in. The example I gave follow the following event impacts:

1. Single disk failure on system 1,
no immediate impact, time to failover to system 2 at leisure. NO DATA LOSS
2. Double disk failure /Full system failure on sysstem 1
immediate failover to system 2 NO DATA LOSS
3. Full OVH Collapse (Disaster Recovery)
30mins to failover DNS external.
10 mins to manually failover to system 3
Up To 10 mins DATA LOSS

None of the above relies on ANY Manual action by SYS of any kind only their automated failover IP system (and we have a manual backup solution to THAT failing!)
Hi john_gb,

I think we are saying the same thing, just in slightly different ways...

I understand your system, but my post was targeted at anyone who may be fearing the worst with these servers. While I agree YOU can design and implement a system with a high level of fault tolerance, this is not possible for everyone. The two main points to consider here are:

Cost. While your system includes 3 servers, not everyone on a shoe-string budget can afford that. Especially if they are doing a project for "fun".

Ability. Not everyone has your level of skill. By way of an example, our Name Servers (yes we run three, soon to be four) are set to 1 hour. Most of the users of SYS servers will use Name Servers from the place they got the domain. For you and me, a quick change is possible, but for others it could well take "learning + 24 to 48 hours".

The main point was to get people to think about how serious downtime would be, and consider how much they are willing to spend. (Please remember there is a wide range beyond what you and I do.) There is a balance to be struck here. My guess is for most people a day of downtime is not the end of the world. The moment you change to "downtime costs me money" then you need to consider a system such as you describe (or indeed mine which is similar). If people are not technically capable, then there are members here (you and I for a start) who can implement and advise...

The other main point of my post was about the decisions which are sometimes made at OVH. The particular example I used was about Oles and the "VAC" Anti-DDoS system. Oles decided to do major work on the network, without telling anyone (for 3 months). I did not know about this until one of my servers was effectively offline. It was not the most important server in the world. It was not the most important server in my world. It could have been though.

The "other provider" (I mentioned in the PM) tell me every time they even think about a network change. I get emails almost daily telling me what IP ranges will be affected, for how long, and what the likely impact will be. (Most often this is "there is a small chance of packet loss or increased latency.)

What I a saying here is:

1. Understand the relationship between price and service.
2. Don't whine when you pay peanuts and get monkeys.
3. Oles, stop ***king with the network.
4. Oles, don't whine when you offer half-price / same-product and everyone moves.
5. Work out what you really want, and if SYS covers it, then buy SYS...

Lastly, I think you missed my point about sleep. If Oles does screw the network at 22:00 tonight, and you are down the pub (and too jolly on your return home to check email) what happens between there and 09:00...? If availability is that important to you, then you need someone (or something) to be on the lookout, and ready to wake you (pi$$ed or not). This happened to me one Sunday afternoon about 2 years ago. I was happily doing the BBQ in the garden, and an irate IT person from one of my customers called (irate as his BBQ had been interrupted as well). Now I get a text if that server is offline for more than 10 minutes (this service costs the client more, but "it is that important to them" - which the the point of the my post...?).

~Trap

john_gb
15-12-2014, 23:33
Quote Originally Posted by mike_
What a ridiculous post.

1. Support shouldn't be poor in the first place
I thought i'd made it clear I was not supporting in any way shape or form the disgraceful support levels! I was just pointing out that you can desiign out dependancy on 3rd party actions!
2. OVH/SYS/KS support used to be good, so a lot of people would't have planned for bad support and ended up getting burned by how dreadful it's become recently
Any tech architect looking at a hosted system would/should have identified the hosting failure events as part if his/her design to achieve the non-functional requirements. Relying on 3rd party actions for business continuity is commercial suicide. The distinction between good and terrible support is NOT one that you any business should be incorporating in their technical architecture, but it is one to consider commercially throughout the lifetime of the relationship, because it impacts the costs you incur.
3. You're running a business on SYS servers?
No, i'm running SEVERAL businesses on SYS, with considerably more redundancy than a lot of businesses over at OVH! How? Because lower costs let you purchase more redundancy!
PS: Using words like 'ridiculous' is very antagonistic, if you have a cogent argument to make then make it, but please show some respect.

john_gb
15-12-2014, 23:19
Quote Originally Posted by Trapper
I agree mostly with what john_gb has said... "Have a Plan" to "....Prevent Pathetic Performance.". How much that plan costs you depends on how important your server is:

If you can bare to be down for a week, just get a SYS server, and be prepared to wait. I am sure the are plenty of uses for a server which only require it to work now and again (they are just not my uses)...

If you can bare to be down for a day, then just order a new server if your one fails. Yes, this plan might mean losing 3-weeks-money on that rare occasion that you have a problem, but it works out cheaper than almost anything else.

If you can only bare to be down for a hour, then Use 2 SYS servers, (or 1 SYS and one at another provider). OR use OVH themselves. This obviously costs more, but you are after something better.

If an hour would be far too long, then you really are in the wrong place. As you no-doubt sleep sometimes, you need an automated system of fail-over. Consider OVH, but also consider other providers.

I also agree with some of the other voices here. If anything takes support longer than a week, there is a serious issue. Even the most minimal of support should be actioned within a week, should it not?

Like many others here I have been bitten by the support problems. Do not limit yourself to thinking about hard-drives either. I had a server which was effectively offline for about 5 days while Oles tinkered with "VAC". There was no warning, there were no statements, there was no testing in an offline environment. He just decided one morning to install a very disruptive piece of hardware on a live network, and then start playing with it...

Thanks to Oles and the VAC, I now use three other providers, as well as "OVH-OVH" and "KS-OVH".

Have a real serious conversation with yourself about how important your server really is. At first, you'll probably say it is ultra-important, but maybe you can convince yourself that actually it is not quite important enough to spend "that" amount of money. When you balance "how-much" with "how-important", then you will have your answer.

Hope this helps,

~Trap
Thanks for the support Trapper, but with respect, you have missed the essence of my post, which is, with the right planning and design you CAN have automatic failover and disaster recovery built in. The example I gave follow the following event impacts:

1. Single disk failure on system 1,
no immediate impact, time to failover to system 2 at leisure. NO DATA LOSS
2. Double disk failure /Full system failure on sysstem 1
immediate failover to system 2 NO DATA LOSS
3. Full OVH Collapse (Disaster Recovery)
30mins to failover DNS external.
10 mins to manually failover to system 3
Up To 10 mins DATA LOSS

None of the above relies on ANY Manual action by SYS of any kind only their automated failover IP system (and we have a manual backup solution to THAT failing!)

Trapper
15-12-2014, 09:20
I agree mostly with what john_gb has said... "Have a Plan" to "....Prevent Pathetic Performance.". How much that plan costs you depends on how important your server is:

If you can bare to be down for a week, just get a SYS server, and be prepared to wait. I am sure the are plenty of uses for a server which only require it to work now and again (they are just not my uses)...

If you can bare to be down for a day, then just order a new server if your one fails. Yes, this plan might mean losing 3-weeks-money on that rare occasion that you have a problem, but it works out cheaper than almost anything else.

If you can only bare to be down for a hour, then Use 2 SYS servers, (or 1 SYS and one at another provider). OR use OVH themselves. This obviously costs more, but you are after something better.

If an hour would be far too long, then you really are in the wrong place. As you no-doubt sleep sometimes, you need an automated system of fail-over. Consider OVH, but also consider other providers.

I also agree with some of the other voices here. If anything takes support longer than a week, there is a serious issue. Even the most minimal of support should be actioned within a week, should it not?

Like many others here I have been bitten by the support problems. Do not limit yourself to thinking about hard-drives either. I had a server which was effectively offline for about 5 days while Oles tinkered with "VAC". There was no warning, there were no statements, there was no testing in an offline environment. He just decided one morning to install a very disruptive piece of hardware on a live network, and then start playing with it...

Thanks to Oles and the VAC, I now use three other providers, as well as "OVH-OVH" and "KS-OVH".

Have a real serious conversation with yourself about how important your server really is. At first, you'll probably say it is ultra-important, but maybe you can convince yourself that actually it is not quite important enough to spend "that" amount of money. When you balance "how-much" with "how-important", then you will have your answer.

Hope this helps,

~Trap

LawsHosting
14-12-2014, 16:31
Quote Originally Posted by john_gb
i'm certainly not excusing OVH/SYS support for poor performance, but if you know that support is poor, then you can plan for it! Lets face it, at the prices SYS do, most organisations can afford to have a spare system just on standby!
If organisations can spare a few hundred or a few grand on servers and redundancy, then they wouldn't even use OVH at all! No?

mike_
14-12-2014, 09:07
What a ridiculous post.

1. Support shouldn't be poor in the first place
2. OVH/SYS/KS support used to be good, so a lot of people would't have planned for bad support and ended up getting burned by how dreadful it's become recently
3. You're running a business on SYS servers?

john_gb
14-12-2014, 06:23
I'm probably going against the grain of opinion here, and i'm certainly not excusing OVH/SYS support for poor performance, but if you know that support is poor, then you can plan for it! Lets face it, at the prices SYS do, most organisations can afford to have a spare system just on standby!
For example, instead of having 1 standalone bare metal server I run 2 virtualisation servers, each cross populates the other with block delta's from virtual machines on a synchronous basis, so a whole system could go down with zero data loss, additionally I asynchronously stream the block delta's to a third server back at base which means that even if ovh got nuked we should not lose more than 10 minutes of data. Additionally we FTP backup regular o/s images to the free backup area, so we can reconstruct a server build in a matter of minutes using the rescue console. All this on 2 x £25/month servers!! Whats the old phrase "Piss Poor Planning...."?

alex
13-12-2014, 23:50
What to do if there is an emergency?

Pray that someone will be bothered about your issue, much easier and quicker to get tickets and travel to France yourself to replace hard drives instead of using technical support.

Maki187
13-12-2014, 21:56
Actualy, u could wait a lot more then 3 weeks if u don't get lucky.

I have something to recommend to you.

Never use default (main) server IP what u get with server, for anything... Emidietly buy 1 Failover IP, and use it instead of main IP, so in case of hardware failure u can buy new server, configure it, move data and just move that failover ip, and there it is, all services working, as u keept IP.
I learned this on bad way.

That will be fastest "fix" of your hardware failure, just to leave old server, or u can try contact support and:

wait a week (or weeks) for first resposne, where they ask u for smart logs (in case of HDD failure), then to wait another week (s) to they ask u did u made backup, and then on the end to wait another week (s) to they replace hardware

Ali
13-12-2014, 21:33
Wow! All these comments are scared the sh*t out of me guys. I have never imaged OVH to be like that. Looks really trustworthy from outside.

I would like to thank for every single person who commented on this.
I'm not sure what to do now. Should I regret to get a server from SoYouStart and try to migrate it to elsewhere asap? I can't stand to have 3 weeks downtime in-case of hardware failure

mike_
13-12-2014, 19:44
Quote Originally Posted by heise
In another case a payment has be set to manual verification and they were not fast enough to verify the payment. Consequence the server got recycled (=everything deleted). http://forum.ovh.co.uk/showthread.ph...ll=1#post66980
And they wouldn't refund the setup fee even though they wiped it just a month after I'd bought it. Had to do a chargeback. theatheist had to fight them to get a refund for the 3 weeks his server was unusable. The company simply hasn't heard of "the customer is always right" or "goodwill gestures". They are incredibly stubborn and if you ever have an issue for which you should be compensated, be prepared to fight and force the issue as they're very unlikely to do it voluntarily.

heise
13-12-2014, 14:40
Hi Danny,

did you ever read http://forum.kimsufi.com/forumdispla...dware-Incident ? Unlike SYS or OVH in Kimsufi they wouldn't even open a ticket if you didn't provide full hardware disclosure yet you can read that sometimes it took 1, 2 even up to 8 weeks after reporting and opening a ticket to get a HDD replaced. Only middle of June until middle of July there was a next day support, while Fabian was taking care of the English section of the hardware incidents. I don't know how he did that.

One customer told me privately, that a technician at your datacenter that was suppose to change a HDD in his server but left in August 2014 and that was the reason why it took over a months to get his HDD replaced.

As everyone can read in http://forum.ovh.co.uk/showthread.ph...is-server-ASAP

09:30 TomOVH - We are working on the disk replacement at the moment
15:32 TomOVH - I have just heard back from my colleague, the intervention is now planned so the disk should be replaced within the next few hours.
17:02 Daniel OVH - Hi, the disk was changed today 16:49 French time

7 hours from the time they are "working" on disk replacement to actually having it replaced and that after over 3 weeks of having the incident reported.

And now the story before http://forum.ovh.co.uk/showthread.ph...6151#post66151

12-11-2014, 12:19 TomOVH - Apologies for the delay in getting back to you, I will take care of this as soon as possible.
17-11-2014, 16:26 WillOVH - You submitted your test results to the validation work queue. I have processed your information and created an intervention reqeust. We will respond directly via email with the next update.
24-11-2014, 18:02 davidovh - The technicians are currently investigating the issue with your server. You will receive an update via email regarding this.



Please OVH were are not against you, we are your customers. Please improve your support.



A lot of time time, the hard drive replacement is delayed due to incomplete information.

We would require SMART logs for BOTH hard drives even if only one disk is defective, we would also request the customer to confirm they have a backup, prior to replacing the disk.

If SYS wants a incident report like Kimsufi wanted http://forum.kimsufi.com/showthread....45-How-to-post why don't you clearly write it somewhere? Besides letting your technician run those commands takes him only a minute.

For HDD problems you can use this from unofficial FAQ:

- What are the commands to run to report a HDD problem?
While this should be easy as the "how to post" explains it, people don't seem to be able to follow it. FIRST don't think which information is necessary to report your problem. SECOND follow all tests to the letter. THIRD copy paste the result into your post. No screenshots. FOURTH tell them they can change the HDD right away.

  1. Boot into (linux) rescue mode. OFFICIAL INFO: http://help.ovh.co.uk/RescueMode
  2. fdisk -l #see all harddisks available
  3. smartctl -a -d ata /dev/sda
  4. smartctl -a -d ata /dev/sdb # if two or more HDD in system
  5. smartctl -a -d ata /dev/sdc # if three or more HDD in system
  6. smartctl -a -d ata /dev/sdd # if four or more HDD in system
  7. smartctl -a -d ata /dev/sde # if five or more HDD in system
  8. smartctl -a -d ata /dev/sdf # if six HDD in system
  9. cat /proc/mdstat # if RAID is present
  10. hdparm -tT /dev/sda
  11. hdparm -tT /dev/sdb # if two or more HDD in system
  12. hdparm -tT /dev/sdc # if three or more HDD in system
  13. hdparm -tT /dev/sdd # if four or more HDD in system
  14. hdparm -tT /dev/sde # if five or more HDD in system
  15. hdparm -tT /dev/sdf # if six HDD in system (they exist)

If you ssh (linux) into your server, you can just copy & paste these commands to the terminal, run the commands, and copy everything back as "CODE" to your new post.


If you should get an error message like this:
The text that you have entered is too long (31984 characters). Please shorten it to 30000 characters long.
then copy&paste half into one post and the other half into a second post.

Dani
13-12-2014, 12:21
Hi

I think i should clarify on the procedure in the event of an emergency.

SYS

The support team for SYS operates Monday to Friday 9-6.
You would look to contact the SYS team via an ticket where by the SYS team will look to resolve the issue via email.

OVH

The support team for OVH operates Monday to Friday 9-6.
You would look to contact the OVH team via an ticket where by the OVH team will look to resolve the issue via email.
You can look to create an incident ticket directly to the datacenter yourself within your manager.
You can call the 24 hour incident number on: 0203 384 2356. https://www.ovh.co.uk/support/submit...nt_details.xml
The option for VIP Support at: https://www.ovh.co.uk/support/vip_su...nformation.xml

---------------------------------------------------------------------------------

In terms of an defective hard drive. SYS and OVH would look to provide you with an replacement hard drive. A lot of time time, the hard drive replacement is delayed due to incomplete information.

We would require SMART logs for BOTH hard drives even if only one disk is defective, we would also request the customer to confirm they have a backup, prior to replacing the disk. Once customers have provided us with all the information, the datacenter staff will look to replace the disk.

I hope this helps, if anything is not clear, please let me know. I will look to explain it.

Danny

theatheist
13-12-2014, 02:25
Quote Originally Posted by Dani
Hi

The datacenter operates 24 hours a day, 7 days a week. If there is an issue with your server, one of the datacenter staff will look to review the issue.

In the event of an emergency, you may look to contact the 24 hour incident number. The number is: 0203 384 2356.

Danny
That is an extremely misleading thing to say. The 24/7 monitoring does not monitor hard drives and it took weeks to get a response from support email and over 3 weeks just to get a drive replaced.

LawsHosting
13-12-2014, 01:34
So the moral of the story is.....pray to hell your server(s) doesn't develop a fault, or more drastically, a faulty drive!

heise
12-12-2014, 23:18
You already guessed it, there are different teams UK, IE, FR, CA and only where you signed up, can help you. People having different accounts often find themselves complaining to the wrong people. http://forum.ovh.co.uk/showthread.ph...ll=1#post66672

Further they may have 24/7 team on site, but that does not mean, that they will fix your problem right away. See http://forum.ovh.co.uk/showthread.ph...is-server-ASAP

Someone already waiting weeks to get HDD replaced!! Then he asks for refund on 27.11.2014. HDD gets replaced 29.11.2014 and the "financial compensation" is only a "commercial gesture". And they close the tread.

In another case a payment has be set to manual verification and they were not fast enough to verify the payment. Consequence the server got recycled (=everything deleted). http://forum.ovh.co.uk/showthread.ph...ll=1#post66980

As far as I can see, the US/CA team offers the best support (=shortest reaction time) on Kimsufi/SoYouStart.

mike_
12-12-2014, 23:04
You don't want to have an emergency on a SYS server. People have waited weeks for replacement hardware.

Ali
12-12-2014, 21:24
Hello Dani,
Thanks for your answer. Nice to know that you guys provide emergency phone number. Can you kindly confirm that it does also cover soyoustart customers?

I, also, haven't been able to find incident report section in my soyoustart control panel. Does it only exist on OVH.ie?

and my last question is that does OVH/SoYouStart UK team also covers requests that goes to Ireland?

Many thanks

Dani
12-12-2014, 21:04
Hi

The datacenter operates 24 hours a day, 7 days a week. If there is an issue with your server, one of the datacenter staff will look to review the issue.

In the event of an emergency, you may look to contact the 24 hour incident number. The number is: 0203 384 2356.

Danny

Ali
12-12-2014, 20:59
Hello Everyone,

My friend recommended soyoustart/ovh to me and I signed up for 2 dedicated server. I've been using them for about 2 months and I'm pretty happy with them so far. I am a system administrator myself so I don't require any kind of technical support and I like easy to use control panel that soyoustart offers.

but I was wondering that what do I need to do in-case of emergency? Does SoYouStart provides emergency support? Seems like their support is really slow. I sent them an email regarding my tax thing and it took over 7 days to get a response. It is totally unacceptable under any kind of emergency situation (such as hardware failure or network problem)

Please kindly advise.