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.
CANNOT run MySQL on a 128MB OpenVZ?
This discussion has been closed.
Comments
Ask your host to allocate you extra 256MB of ram for 24 hours, most will be happy to do so.
I'm using 128mb boxes for my clients' low traffic wordpress sites, one site per box, seems good. The sites' load speed is much faster than shared hosting.
I'd like each site has a IPv4 address. If I mess up, it would only affect one site.
There are many attacks targeting wordpress sites so I'm still looking for the best backup practice. ( no backup server information stored on live server, easy restore, easy to set frequency and store version )
Now I want to do it like this:
live server-->backupserver-1-->backupserver-2
1,tar files and database
2,move compressed file to a directory
3,set a non-exist domain server_name to that directory
4,modify hosts on backupserver-1, point domain to IP of live server
5,wget file from live server everyday, store 7 days file
6,do the same trick on backupserver-2, wget file from backupserver-1 every week, store 4 weeks
@serverian @Maounique
Do you think it would be a good idea if I host more sites on a bigger box with multi IPv4?
This seems an unjustified waste of IP space.
WTF?
Y'know there are services where you take your backup once and you pay other people to worry about making numerous safe copies, right? They're not back bracingly expensive either.
I hope not... last I heard LEA was alive and well.
Times have changed since LEA did that famous 64MB experiment. Well, mostly MySQL has changed -- the default storage engine to InnoDB, and some other parameters that make it harder to run it on a true LEB. Greedy devs with no low end spirit
Code it yourself and move to mongo... you know you want to
A real LowEndBoxer uses SQLite... and codes it (him|her)self
But then how do you shard your database over all your low end boxes (and end u p with worse performance than one box)?!
A real LowEndBoxer only needs one lowendbox
I wouldn't say I need any of my boxes, I still have 8...
Ahhh, you also have LowEndBoxia, the tendency to collect lowendboxes for no reason.
I have more than that, but I'm currently clearing out the clutter, the ones that aren't production-worthy: DeadicatedMinds, WeDontLoveServers, etc.
We started to offer shared hosting with dedicated ipv4 by default, so one misbehaving customer wont affect others, especially resellers which host any spammer and their dog. Spam goes out, service suspended until the IP is clean again, simple.
Yes and no, with shared hosting you have an extremely high chance to get someone unable to update their wp or gallery script and this ends up badly most of the time. We have it often enough on zPanel VPSe which need a lot more skill than simply putting up an wordpress with softaculous.
I'm actually running a ~2.5GB MySQL database on a 128MB VPS, but with an SSD disk.
Buffering isn't the best (not too bad though, as most of the data set is rarely read) but the SSD makes it a non-issue, I've found. I've been thinking about whether it makes more sense to have more memory, but at $15/year, the most you can get is around 512MB.
If I were to cache the entire DB to memory, I'd probably want at least 4GB RAM (the DB is constantly growing in size after all) which would be significantly more expensive than what I'm paying for right now.
Perhaps more RAM, minus SSD, is the way to go, however I like the consistent low latency reads I get from the SSD as opposed to the greatly variable latencies I often get from VPSes using spinning disks.
Memory caching only really helps with reads though - writes still need to hit the disk and don't cache nearly as well.
That is mostly ok with ssds however, I have seen cases in which read/write remain over 10 MB/s non-stop reaching 200 MB/s in cases when KVM is involved and starts swapping. That is abuse and unless the node is mostly idle, cannot be left unchecked for long.
I totally see your point, although this is probably more the case of poor configuration causing the VPS to swap, due to database buffers being too large.
Not really/always, it can work fine for months and then suddenly start doing so, when the ram is so low any mistake or bad luck causes severe malfunctions. It is basically playing with fire and investing way too much time in supervising/tweaking than it is worth to save haf a dollar. It is simply not worth it.
Fair point, but mistakes, bad luck or severe malfunctions can happen regardless of the amount of RAM.
Each to their own opinion though.
You are a butcher.
I also had problem before with them on mysql 128mb openvz, I was a noob before ( untill now) and with their policy "this is unmanaged service go fix your shit yourself", I just decided to switch provider who's willing to extend a bit help, then problem solved.
I prefer the term "fresh meat provider". =P
You could run a 1TB database on a 128MB VPS if you were the only one using it and your queries were simple.
On the other hand, 1000 concurrent users of a 100MB database could bring a 512MB VPS to its knees.
I will second to that
Quite aware of that, as well as the other corner cases you could draw really.
I get around 1 million queries per day on this box - not a huge amount, but certainly more than the blog that only gets visited twice a day, and not unusually less than compared to a typical website you'd host on a 128MB RAM server. Most queries are fairly simple, some aren't, but I expect that from a typical application.
As for your second example, assuming good configuration and typical queries, it's unlikely that the disk or memory would be the cause for problems. I know you could concoct situations where this doesn't apply, but I'm speaking from a typical usage perspective.
I've had mysql installed on a 32GB ipxcore vps as part of lamp (with php5-fpm), although not used much.
Locales would not install however. I installed locales using sshfs from my desktop to mount the vps's disk then chrooting into it and then installing. Slow but it works.
It depends on the application written. If it's something retrieving only by primary key with seldom writes, 128mb can go a long way.
Boleh minta confignya? Kok keren
bro