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
One must always do backups, even with the greatest host on Earth
i did backup, always! just lazy to rebuild vps
Correct me if i'm wrong but all this disaster happened during MC doing upgrade? Not natural earth quake disaster right?
Hey, look all major vps provider give proper email notification and time frame before doing maintenance.
If only he inform ALL clients and give time range when upgrade/fixing things executed... Could save other people who don't do regular backup to prepare the worst thing like this.
my 2 cent ;p
Due to the way things marry up, between cache and spinning disks. Either part of that equation going south means a big problem.
I've heard that Solus may be redoing the backup system in the future, if so, when and if that happens I will look at deploying backups for VM's again.
Extremely sorry for the trouble. Working non stop to get things back in working order.
I am a customer of MCH. So all data was lost? Did Phil do backup (at least monthly)?
If only you give customer atleast 2-3 days before doing anything at can damage system to backup....
Do what i do and run cdp.me + the DB backup
mine & billing (whmcs) still down
Taking backups is yours and only your responsibility. It clearly says so to TOS/AUP of most LET providers. And it has been said here a trillion times. If you buy an unmanaged server, then, you HAVE to know how to use and maintain it. This includes taking regular backups in another online and, even better, also in an off line location. As more freequent your backups will be, as better for you in a case of disaster.
TL'DR Do not blame provider for not taking backups, he has not an obligation to do it, except if it is written in the terms of service.
If MCH leave a time window of 2-3 days applying an emergency vulnerability patch, just to give his customers "time" to do backups, then, he would be a terrible provider because he would leave his clients for 2-3 days open to the exploit...
Not within dayS, can be minuteS or hourS, at least inform/email ALL clients before execution.
I got my backup as always and not-so-panic-condition when this down happened, but i blame MCH for doing not aware this customer-provider relationship. He take action on the first place without notifying clients, so its his responsibility.
Its been long run and very pleased with this provider, great services but this action caused BIG trouble. Lesson learn...
Just move on
I quite sure @MCHPhil will settle your issues, just need to be patient. These things take time, you know......
Actually he did notify clients before the maintenance via email. You said you received one of the update emails above, so he's sending them to you too. Perhaps the original email landed in your spam box?
Nope, not received the first mail, I check regularly my spam inbox lol
looks like OP @FredQc also not receiving the first email
so is this one of the downfalls of running raid0 in a hosting environment ?
yes. they should learn from Vultr, who as we all know don't use any RAID and have never lost customer data. /s
Canada shared hosting is still down
@rokok This is the first email. It has been send 3 hours before the incident. Maybe it is your issue (reject emails, spam filtering, other deliveribility issues) that caused a non-receipt.
Issues does happen even in the largest providers. It limits @MCHPhil 's responsibility just in frame as described on TOS/AUP:
We are not liable for any loss of data for any reason.http://web.archive.org/web/20150315005625/http://mycustomhosting.net/tos.php
Also, Phil in the second email is doing his best to satisfy customers that did experience downtime:
If you don't like this attitude by a provider, move on. But, I highly dought that you will find such so, in the entire LET/LEB or even more expensive ones that offer unmanaged services...
It looks like.
SSD Cache issue shoulnd't ever be related to DATA loss. You can rebuild the ssd cache (if flashcade) any time, without data loss. So, definitely, raid 0, it seems.
Not gonna comment anymore, i paid yearly, not gonna do anything
Im already move on
In that case I hope he wasn't joking about shared hosting having no dataloss.
from previous offer postings he said the KVMs resided on RAID0 + SSD Cache , so I guess his luck finally ran out on raid0.
Oh, the second time data loss occurs at MCH, guess Phil learned nothing the last time it happened.
Majority had nothing to do with RAID 0. CAVPS4, yes. 2,3 no. I do not use a widely known or used cache system. I do not speak much about it and do not plan to start now.
The way my cache system works, as stated, if they are not brought offline correctly, they are toast.
not like we are established or anything but if you are running mission critical stuff and need something to fall back on, let me know. location CA, OVH BHS.
that being said, I'm quite sure mycustomhosting will have you back online soon.
When you buy an unmanaged vps or dedi, you have to know what are you buying. That means, you have to search the specs of the node or the vm (cpu, mem type, raid type, storage type, network etc.) and then, decide if it fits for you or not.
@MCHPhil 's product is RAID0. That means, good performance with (theoretically) a higher possibility of disk failure (well, I've seen some providers with raid10 lose all of their node's data because of failure of the hardware raid card...).
So, in a vps costs 6$ per month with 3GB Ram, 30 GB SSD cached space and 3TB traffic, you have to chose not to rely exclusively on the best realibility of this vm, exclusively on not-failing-HDDs. That means you should have backups and, if the service is critical, a way to restore it to a second server immediatelly (e.g. a mirror setup with continously rsyncs and mysql replication or similar). If it is even more critical than the uptime of a website, then, you should not buy a 6& per month service to rely on.
There is e reason that some services cost x20 or x30 than those vps's. It is not so much of performance (I've seen better performance from some LET providers than in an Amazone same specs vm) but it is a matter of realibility and methods to ensure that in any case of a catastrophy, there are several setups that will keep you online immediatelly.
+1. @MCHPhil provided the service he sold, nothing more and nothing less. And I still think his service is awesome for the price.
Unfortunate... but it's nice to see he's posting and emailing updates! Appreciate upfront providers like that, even though it's crazy trying to get things back online...
You are NOT right.. there is a major KVM / XEN HVM security bug announced yesterday http://lowendtalk.com/discussion/52536/extremely-serious-virtual-machine-bug-threatens-cloud-providers-everywhere#latest
if your other VM providers already do not do emergency shutdowns or announce that they give you 2-3 days they live you vulnerable for that period...
correct decision is emergency shutdown -r now....
P.S I'll not comment RAID 0 issues but that is stupid....
Why were nodes rebooted? This security vulnerability only required a reboot of customer VMs to be patched.
Raid 0 + a write back cache?
That is just irresponsible.
You are right, but unfortunately, we can't always trust customers to follow the reboot email. So most of us take the opportunity to reboot the host also.