New on LowEndTalk? Please Register and read our Community Rules.
All new Registrations are manually reviewed and approved, so a short delay after registration may occur before your account becomes active.
All new Registrations are manually reviewed and approved, so a short delay after registration may occur before your account becomes active.
MyCustomHosting Down ? (Resolved. Issues? Open a ticket first!)
This discussion has been closed.
Comments
Mcphill just having a bad day, best of luck he can handle the situation and not ended like GoodHosting.co lulz
Hmm I'm missing GoodHosting, let me bump the thread
Boo support site still down
Whether or not any provider claims to hold backups, it should be your responsibility as the customer to back up critical data. If it's that important to you, then you should be doing routine backups to other hosts or at least keep an offline copy to restore from as a last resort.
Shit happens, and for $7/month you can't complain too much. LET offers are cheaper because they are seen as an 'unmanaged' service. You want a provider that schedules offsite backups for you? Feel free to spend $50+ per month getting a fully managed service. Otherwise, please quit making a huge fuss because you were not responsible enough to back up your critical data.
OK then, go and buy things from unknown providers, don't look for the specs of the node but just rely on phrases like "we are the best hosting company in the whole world", don't look after what other customers have been said in forums.
If this seem joke for you, I don;t think so. When I want to buy hosting for bussiness purposes, then, I search if the provider is new to me and I look what the specs he offers me are, to decide if it will do the job.
@jvnadr you're self-contradictory
Shit happens, all the best on recovery.
How? Where do you disagree? On searching before buying?
@jvnadr first, you said people should go around and search for the node specs, because providers will not tell you if their setup is lack of something. Maybe those providers will be good for you, not me.
And MCH does describe his nodes setup, and has been answered here to several threads about the configuration of his servers stating that he uses raid0.
Then you said providers should let customers know about the specs
Or my English is too bad and there is misunderstanding here?
@comXyz Do not put words in my mouth. I never said that providers will not tell you their setup. I said that providers will NOT ADVERTIZE the specs that are not so good. They will tell you, they probably will write them in their site (with small prints, if they don't is a minus) but they will not get public shouting about them. THey will shout about the good characteristics, not the others.
If you see an add of a fridge that is big enough but eats a lot of energy, the add will say "you can put in the fridge tons of food". They won't advertize that it will cost you $$$ in your electric bill. But, if you search the specs in the manual, you will see exactly how much energy this fridge needs to operate.
So, if you go to a web site to order this fridge, then, you should not only look what the production company is advertising, but search the manual, the complete specs and, if this is possible, search in forums / web what other people that have bought the same fridge saying about it. Are they happy? Did they find out any problems?
If you just order a fridge because you saw in an add that it is... gorgeous, do not cry when it comes home and realize that there was some specs you didn't like them. Search first.
Hope I am clear to you now...
And you're self-contradictory again.
OK, I quit...
So... don't know if anyone can trust this provider again. Will the servers go poof next time they need to be rebooted or if they lose power ? I don't recall reading anything about configuration changes to prevent this from happening again.
well, standard reboots (when necessary) are a norm, should never ever cause server issues, not even with software patches.
will backup config files on github for quick restore of VMs whilst on the road. had data backup scheduled on daily basis. so no big deal here.
Just created a fullbackup of my site, few hours of dataloss but no biggie, moving it over to my buyvm reseller.
I like to take a moment to thank @Anthonysmith. You were the first one to say I was using a write back cache, but later forgot you made that up? I would extremely appreciate it if you didn't start rumors.
MCH does not, has not and has no plans of ever using a write back cache. The RAID controllers offered by OVH do not stack up to the performance I can get without them, using my hybrid setup.
I didn't try to deceive anyone, you can clearly see most customers know that I don't take backups for the KVM line up. (There was a mass mail sent out to inform them of that a while back.) Honestly, who would? It takes over 24 hours to backup a node and almost a week to restore. I don't toss 40-50 VM's on these and say done, these are loaded. It's how my business model works.
If you had any questions about backup it is YOUR duty to ask as a consumer. It is also your duty as a consumer, no matter if a provider has backups or not, to take your own backup. If your data is not in at least 3 places it is not important.
When the decision was made to cut KVM VPS from backups, we still offered OVZ and web hosting which was/is backed up. We dropped OVZ completely. I am at fault for not properly distinguishing this in a few spots on my site, this is corrected. I had originally only removed the backup feature on the KVM VPS page, as that is the page I'd look at if I wanted to know what features the KVM VPS come with but I was wrong.
Lets reverse this here, had there been backups of the KVM VPS, you realize we still would be down. With no end in site. Transferring 16TB of data does not happen instantly. And then with the way SolusVM packages that data, you have to untar/decompress 16 TB of data. And then restore the VM's, so week+ at least.
All VM's were back online within 24 hours and the webserver less than 48 hours.
Shit happens. I'm sorry.
My setup is my setup, you may not like it, but it works well and is fast. It provides me the ability to offer services at the prices I do. Want KVM backups? Expect prices to raise. Want RAID 10? Expect the service to disappear. RAID 5 with 2 spinning disks and 2 SSD? Nope. RAID 1? LOL, I would lose money if I tried and performance would be horrible. Anyone who has ever asked about the hardware has been told straight-up, with no bush beating. You can see that in the many offer threads I have here. If you wanted to know what I use for SSD cache (NOT! a write back cache), toss me a PM. Something I've stated multiple times also.
Hell one guy asked me about my raid setup in multiple offers. Consistently.
In the end, RAID1,10,5 NOT RAID 0 would have save 1 of the servers. The only hint I have so far on the others is an issue with iptables not unloading properly and causing the machine to hang. Hard to troubleshoot something that hangs and eats data. At this point the host nodes are not utilizing the SSD cache. VM's are. This should help.
If you would like to offer something else out of BHS OVH, go ahead. I am not stopping you. There are very few providers that do so and that is because of the challenges OVH presents.
The billing system was due for some changes so I will be using this time to facilitate those. Plan to have billing system up and functional by end of the day. As stated, credits will be handled at the end of this fiasco. Anyone with an open support issue with be triaged and answered accordingly, expect delays.
Thank you for your support and understanding!
Have 2 VPS' with MCH for over a year now and up until this happened they've been 100% rock solid and have actually worked as my fallback servers when others went down due to various issues.
It sucks this happened but due to taking proper backups and whatnot, setting up both of my servers took me a total of 1-hour combined and all is well again. The only thing that is truly gone is a bit of time but oh well, shit happens...
Best of luck to Phil and hopefully everything gets up and running ASAP and he's able to learn from this and improve MCH even more.
@MCHPhil there is no other sort of cache that would cause this unless you actually cached the root partition to and were using LVM with thin provisioning and did not back up the metadata, as you wont say what you use that is the only logical assumption based on actual experience.
quick edit: just to say if your providing performance VPS which seems to be the case that is fair enough, its a good model but the DR plan clearly needs to be considered.
@MCHPhil
So, in your signature you state that transparency is the key, while, if you read a few posts above, you mentioned that it's your own setup and you wouldn't talk about it further? ....
What kind of SSD Caching are you currently using?
@MCHPhil ok the auto network config hasn't worked and I can't access the knowledge base; can you please post the manual instructions somewhere so I can get back online?
>
What's the difference between sending you a PM and posting it?
I've heard you're a nice guy, though, I'm not attacking you, at all.
Making it readily available for some one at any point to come and use. Telling one person is different. But whatever at this point. I'm tired of the berating, I will say it here and people will still not know and continue to ask for 10 pages. And due to @anthonysmith it will be remembered for a lot longer that I FALSELY use a write back cache. I use tier.
Phil, that's what transparency means.
We've had to reboot our node multiple times for testing, currently using flashcache and had no data loss, although we run RAID10 + SSD.
As for AnthonySmith's post, he simply assumed due to the data loss you suffered, he tried to get an explanation.
A normal reboot would have been fine. See CAVPS5 had no issue. The problem is the two machines hung. Frozen. After 3.5 hours what would you do, power cycle and hope things are ok. Your RAID10 does nothing if the FS is destroyed.
I don't think anyone would want to copy that setup after the recent events
Anyway, until now had no issue at all. I had backups even if my VPS was untouched and the files were there.
I continue to use your services as I do for many months already.
I don't understand the problem here. As far as I know @MCHPhil was always open about his setup; I seem to recall him posting about it in other threads. Would a different RAID setup help? Maybe. Certainly it would mitigate some types of failures but RAID is not a backup and will not protect against file system corruption. I lost my VPS data as well but I only used it as a dev/test machine--so while I didn't have a backup, I also don't really care that the data is gone.
It's pretty obvious that Phil has worked diligently to resolve the situation. Most providers do not guarantee backups and even if they do users should have their own backups. If uptime is important data should be replicated across multiple VPSs in multiple locations like some have mentioned.
Is it disappointing? Yes. But it happens. There's no surety to any of this (RAID 10 can also fail).
nothing to express except lol. What do you except more for a cheap service ?
a backup ? come on, be realistic.