The site will be offline for an hour or two today. We're migrating to a new server. Search may be down for a few hours longer after the site comes back while I get it hooked back up.
Moving to LiquidWeb's newest "zone" with nicer tools.
Upgrading from mechanical to SSD drive.
Upgrading from CentOS 5 to 7.
Upgrading from Apache 2.2 to 2.4.
Upgrading from PHP 5.4 to 5.6 (big performance gains).
Upgrading from MySQL 5.5 to MariaDB 5.7 (a fork of the project by the founder).
Slightly less $ per month for overall better performance and new tech stack.
Upgrading our web stack will be much easier in the new setup.
Our old web stack was quite fragile, to the point where the techs didn't want to touch it to do a simple upgrade like moving a point version in PHP. So I decided back in August to do a full migration. After a few fits and starts, I sat down to set it up in earnest yesterday. Turns out I've gotten a little faster at this process over the years so I expect we'll be moving as soon as I finish doing a dry run with the databases today.
Moving to *#%^&!Q@( newest "zone" with nicer tools.
Upgrading from mechanical to SSD drive.
Upgrading from %$&^@! 5 to 7. the first number is smaller than the second
Upgrading from &#$ 2.2 to 2.4.
Upgrading from &^& 5.4 to 5.6 (big performance gains).
Upgrading from ^#$( 5.5 to %(@# 5.7 (a fork of the project by the founder).
Slightly less $ per month for overall better performance and new tech stack.
Upgrading our web stack will be much easier in the new setup.
Our old web stack was quite fragile, to the point where the techs didn't want to touch it to do a simple upgrade like moving a point version in PHP. So I decided back in August to do a full migration. After a few fits and starts, I sat down to set it up in earnest yesterday. Turns out I've gotten a little faster at this process over the years so I expect we'll be moving as soon as I finish doing a dry run with the databases today.
@ardichoke said:
You could have called me and I would have gotten it running for you, met you in Detroit and had 4 rounds at the bar in that amount of time.
@Snarkasm said:
Weird, it's like he has enough experience to get paid to do these kinds of things.
Unlike the people I'm actually paying to do it that left me hanging for 2 days thru levels of idiocy I can't recount without raising my voice. Not that I'm salty.
@Snarkasm said:
Weird, it's like he has enough experience to get paid to do these kinds of things.
Unlike the people I'm actually paying to do it that left me hanging for 2 days thru levels of idiocy I can't recount without raising my voice. Not that I'm salty.
To be fair, I get paid a LOT more than those people. There's a reason I was managing internal systems, and not customer systems.
Bit of Icrontic lore / Linux nerdery: Back in the days of bart, our first server, the websites had to be served from a secondary hard drive. I think our system drive was super fast for the time, but super small, and Icrontic was quite large for contemporary hosting standards. So we mounted a (5400rpm?) drive as /storage and that was where our webroot (and databases, and logs) were put. And then I just... left it that way. A bit of paranoia about breaking something obscure, and a bit of laziness. 4 servers later, I've finally moved the webroot under /srv and left the databases and logs where they go by default like a grownup sysadmin.
@Linc, it's nice to do things right, but no one will blame you for operating for a time under the "if it ain't broke" phiosophy. Most systems that have been around for a few hardware, os, and software changes will show some "historical" configuration. Good job getting it all done and thanks for keeping the site working well.
@Linc said:
Bit of Icrontic lore / Linux nerdery: Back in the days of bart, our first server, the websites had to be served from a secondary hard drive. I think our system drive was super fast for the time, but super small, and Icrontic was quite large for contemporary hosting standards. So we mounted a (5400rpm?) drive as /storage and that was where our webroot (and databases, and logs) were put. And then I just... left it that way. A bit of paranoia about breaking something obscure, and a bit of laziness. 4 servers later, I've finally moved the webroot under /srv and left the databases and logs where they go by default like a grownup sysadmin.
/srv isn't the standard location for docroots on RHEL systems either. /var/www is.
Not that it really matters that much. RedHat takes forever to make changes and the Filesystem Hierarchy Standard added /srv 11 years ago.
Comments
Nerd details:
Our old web stack was quite fragile, to the point where the techs didn't want to touch it to do a simple upgrade like moving a point version in PHP. So I decided back in August to do a full migration. After a few fits and starts, I sat down to set it up in earnest yesterday. Turns out I've gotten a little faster at this process over the years so I expect we'll be moving as soon as I finish doing a dry run with the databases today.
but can it run crysis?
No, but it can turn into one.
All sites except Icrontic and NewBuddhist have made the jump now. Going to get Sphinx running & indexed on the new server, then pull the trigger.
thats what i got out of that. cool beans
10 hours later...
You could have called me and I would have gotten it running for you, met you in Detroit and had 4 rounds at the bar in that amount of time.
It still isn't working.
fml I let @ardichoke take a look and he had it fixed in under an hour.
Weird, it's like he has enough experience to get paid to do these kinds of things.
Unlike the people I'm actually paying to do it that left me hanging for 2 days thru levels of idiocy I can't recount without raising my voice. Not that I'm salty.
Final file transfer has commenced. Downtime imminent.
See you on the other side
BAM!
h4x0rs in teh yu0r webserver
Yes thank you for screenshotting my single mistake in that goddamn work-of-art server migration.
Don't worry, I've carefully crafted a ragequit post as a result
I hope it has no fewer than 5 witticisms to make me regret your absence.
Where's the new logo? I was promised a new logo.
Can't see that phrasing and not be reminded of this commercial:
To be fair, I get paid a LOT more than those people. There's a reason I was managing internal systems, and not customer systems.
Bit of Icrontic lore / Linux nerdery: Back in the days of bart, our first server, the websites had to be served from a secondary hard drive. I think our system drive was super fast for the time, but super small, and Icrontic was quite large for contemporary hosting standards. So we mounted a (5400rpm?) drive as
/storageand that was where our webroot (and databases, and logs) were put. And then I just... left it that way. A bit of paranoia about breaking something obscure, and a bit of laziness. 4 servers later, I've finally moved the webroot under/srvand left the databases and logs where they go by default like a grownup sysadmin.@Linc, it's nice to do things right, but no one will blame you for operating for a time under the "if it ain't broke" phiosophy. Most systems that have been around for a few hardware, os, and software changes will show some "historical" configuration. Good job getting it all done and thanks for keeping the site working well.
/srvisn't the standard location for docroots on RHEL systems either./var/wwwis.Not that it really matters that much. RedHat takes forever to make changes and the Filesystem Hierarchy Standard added
/srv11 years ago.Well now I feel old. In my head,
/var/wwwis current. Sigh.It is in mine as well. But I primarily work with RHEL compatible systems, and I stick to their standards whenever possible.
We use Ubuntu at work. Easier for me to remember
/srvsince it's what I use all day.Is CentOS still basically Red Hat?