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.
Comments
Solutions.
My website is not responsible for this. It does not have a large amount of traffic (the logs show GET requests are a few seconds apart for the periods in question, and it doesn't use persistent connection). So I won't even have 2 simultaneous connections, never mind 20 (the limit on my account with Asura). So it must be the total server limit (800) exceeded, and this could possible be due to some DDOS attack on other sites using mysql on the server, in which case DDOS protection is obviously failing (as mentioned this only started on 14 August; previously I never had this issue since I started hosting with Asura)
I am not insistent on it. I just wanted to point you to the possibility of it, given that Asura have been unable to fix this for 2 weeks now, and the fact that you are providing these services to them (which obviously are not fully under Asura's control, certainly not as far as the hardware is concerned).
Didn't say it was. The SQL server you're connecting to has a maximum connection limit that all users on the server are sharing from. It could be a single user absolutely hammering it and causing the issue. Its pretty straight forward to increase this limit, we've had to do it on most of Namecrane's servers as of late.
There's nothing I can do on my end to help you with your issues at Asura.
All the best.
Francisco
Singapore ETA?
This week from the looks of it. They're supposedly finishing up racking/cabling.
Francisco
Any updates on when the NY node will be upgraded?
last time i had ticketed Asura they told me they use OVH servers. (now don't contact ovh to fix this issue LOL). Also in past somewhere I read they were using buyvm's big-fat slices too along with colocation gear and putting shared/reseller hosting plans on it. Even if that was were to be true, still buyvm/frantech or Asura's whomsoever upstream/provider is at no fault of SQL connection's etc because afterall thats "local" feature of the node/server and only Asura can fix this. So blaming the upstream will not help, neither they can peep into client's servers to fix if something is broken unless thats a fully managed service which I don't think Asura has signed up with.
Now, lets not go further to dig-in about Asura in this thread. but if your site is valuable and you think asura can't fix simple sql problem, then why not just migrate it to namecrane? NC Support is good, they fix things super-fast, and with new genoa servers theres no lag/buggy SQL since I myself host many clients with them.
>
Is this done / planned ? Are we expecting any downtime ? I don't remember seeing an email related to this ..
Racks all done, should be migrating in the next hour or two.
The email I pasted at in the OP is what we sent to everyone a few weeks back.
Francisco
Migrations are running
It'll be done in 3 - 4 hours. We'll then have to power off/on to apply the resource changes.
I'll likely do a 2nd outage on the DA node while i'm at it to move SQL back onto the same node. That'll take 10 - 15 minutes.
Francisco
Upgraded!
I gotta move MySQL in a bit, but the rest is all good.
Francisco
lup yu fran
Hi @Francisco, any estimate for when NY will be completed?
When it's ready, please announce it here. Thank you.
Telehouse has been slow getting the rack provisioned. Hopefully soon.
Francisco
Hi @Francisco Is there any issue with the Singapore Namecrane node? Currently, the web server is down, and we also experienced brief server downtime yesterday, along with Redis being down again.
Singapore Namecrane node unstable after the hardware upgrade.
Yes singapore is still bumpy
Any updates on the NY node upgrade @Francisco?
Thank you.
Telehouse finally agreed to provision the cross connects as they signed off. They agreed to do one time cost on them and then tried to change it to monthly recurring after everything g was signed and approved.
It caused a delay and only got cleared up on Friday.
Francisco
which page /app is?
hetrixtools
NY is still running rock solid for me on the 7900. Server load rarely exceeds 10 and uptime of 99.97%. On a 9654 it would be idle lmao
The ryzen's get super busy once everyones backups start firing and things like that. Genoa's are ordered and sitting in the sellers warehouse appreciating in value because of how RAM/NAND are going through the roof.
Francisco
Will there be an upgrade this month?
Hello @Francisco ,
Is there any specific reason why the Namecrane Singapore node restricts sending emails via external SMTP?
Every time I create an account for my client, I have to open a new ticket to whitelist that account so it can send emails through an external SMTP service (such as Mailjet, Resend, etc.).
Is there a particular reason why the Namecrane Singapore node blocks outbound connections to external SMTP servers by default?
Faced this same issue - I had actually whitelisted to Cranemail SMTP, and they blocked it without notice, and I had a few blesta related mails all fail and had to ticket in to whitelist again. Not a good experience sadly.
Honestly this should be the default action on all of our nodes, we just haven't implemented it.
SG got it first because there was a user that was making outgoing connections to remote port 25 destinations (spam, brute, whatever), which got it blacklisted on spamhaus and abuseipdb, which then got malwarebytes and the rest blocking access.
You should just ask your mail hosts for alt ports, everyone offers them.
Francisco
If using SMTP from Namecrane, I can understand that Namecrane might be concerned that the email service from the server could be used for spam, brute force attacks, or other negative activities. However, if external SMTP services are also blocked, isn’t that strange? If an external SMTP service gets blacklisted, it’s that service that should be affected, not Namecrane’s server, right?
And one more thing — the one being blocked by Namecrane isn’t port 25, because I use external SMTP services with port 465 (SSL) or 587 (TLS), and yet these are still blocked by Namecrane by default.
Blocking or disallowing the use of Namecrane’s own server to send emails makes perfect sense, but blocking external SMTP as well is confusing — how will my web app contact customers via email notifications then?
Sure, I could open a ticket to request an account whitelist, but that’s quite inconvenient, especially if I forget to open a ticket or the response is delayed while the web app needs to be used immediately.
Fran, any update on upgrading the NY node? It would be nice to have two CPU threads per subaccount.
If using SMTP from NameCrane, they can monitor all outgoing emails for spam. They have control over it.
The issue is unwanted outgoing connections. There are PHP scripts that can scan IPs for open ports and bruteforce credentials. Using these scripts will get your IP blacklisted, as is what apparently happened on the Singapore node. Once blacklisted, everyone will face issues with their website and email.
At the end of the day, the problem isnt NameCrane. It's an issue on all shared hosting providers. If you want total freedom to connect to any ip:port freely, get your own server. Otherwise, submit a ticket. NameCrane responds to my support tickets with a resolution in usually under 10 minutes.
@Francisco
Are those 4 x 7.68TB NVMe in hardware raid? Or mdadm?
mdadm.
Francisco