It's recommended to make thread_concurrency = Number of CPU's x 2. But this is really effective with latest mysql5 versions?
[url]
Default 10
Range 1-512
Why to set it to 8 (4 cpus), for example, if recommended default value is 10? And why to set such a low value if range is till 512?
Also, how to check assigned value as "show variables;" command omit this value.
Assume that I want 3 URLs -- a main, 'real' URL and two misspelled variants.
Through a domain registrar, I register theCorrectSpelling.com, misspelling1.com and
misspelling2.com.
I retain the services of a web host and set up theCorrectSpelling.com with them.
What's the mosti) effective way
andii) cost-effective way
to establish misspelling1.com and misspelling2.com so that they refer/point/bounce to theCorrectSpelling.com?
Do I have to establish additional sites (with potential associated costs),
what's the correct value for thread_concurrency in /etc/my.cnf if the processor is
Dual Processor Dual Core Opteron 270
is it 4 or 8?
I have a core2duo at SoftLayer. MySQL 4.1.20.
I tried to add thread_concurrency=4 to my.cnf. I restarted mysqld, but thread_concurrency doesn't show up in "show variables".
Also, is TOP supposed to show the loads in each core separately?
setting up the MySql thread_concurrency variable on a Single-Processor Xeon 2.6ghz QuadCore .
Should it be 8 (num of cores * 2) ? or 4?
We are starting to bring a few servers in-house rather than leasing them. We decided to do it ourselves for our email server and a few others. We are starting with a 15mbps commit on fiber (via ethernet hand off), and don't plan to exceed 20-30 any time soon, but if we did we need remove for expansion on equipment.
What Router/Router Series would you guys recommand for a small budget friendly project, but more importantly something that is very stable? I'm pretty tech savvy, however easy to configure and maintain will be high up on the list....
Also what switch brand/series would you recommend that are cost effective and can handle a decent load?
I had a program (APF, actually) install its executable file in /usr/local/sbin/, so I went to add that directory to my PATH variable in .bashrc. I'm able to type commands to executables just fine in other directories, but for some reason I keep getting "command not found errors" from bash when I type in the name of the executable (apf), even when starting/restarting new instances of the shell. Here's my current PATH variable in .bashrc - have I overlooked something?
# Customize the path directory
PATH=/bin:/sbin:/usr/bin:/usr/sbin:/usr/local/sbin:~/bin:
Some of you may have read my previous posts about a dual server configuration I am currently working with. I run a high traffic forum which has up to 2-3k of people online at once. I was wondering if it could be effective to setup MySQL Replication of certain tables which are read very frequently and then modifying the script to grab data from the slave server rather than the master? For say viewing threads, forums etc. Information which isn't updated literally every second.
A few questions...
-Will this place a lot of load on the master having to write the data to the slave as well? As in would the load I save on SELECT queries be used on writing to the slave anyway?
-Could this actually be effective?
i've got a server that averages 3-4 TB a month. it is starting to max out the limits on the servers capacity and i want to setup another server on a different network that can help load balance, and if one of the servers are down for any reason the other server would take on all the load while the other server is down.
what is the best solution for this?
First off about myself: I'm your typical knows-enough-to-get-by web design CMS flash type guy. Nothing special no real training but if I need to run a control panel I probably can.
What I need: I'm moving an AM station(oldies) to online streaming and need a host for it. The variable here is the only reason I'm moving the station is because a radio show on it which runs 1 hour every weekday needs to expand our audience and all that good stuff. We'll be expanding the show as far as we can hopefully within the next year and on certain days we'd have more listeners due to having higher profile guests on.
However, since we have the stream 3 other shows on the radio station have offered to pay for the costs(we had to get a cable line run to the station in the first place)- presumably they would have a lower listener rate.
The question is what's the best solution here? We don't want to pay for a 2000 slot shoutcast service right off the bat, but we also need one who can rapidly expand our slots as needed, or possibly have one which just charges based on use. When we get up to high volume would it be smarter for us to just lease an unmetered dedicated server (high assumption of 1000 listeners(which is months off) for 3 hours a day running at 128k bit-rate the bandwidth really starts to add up) or a dedicated server that just charges based on usage. The latter may even be cheaper than using the high-priced-slot-based shoutcast stream servers.
I figure someone out there has already done the research can might know an answer to the problem. The key here is flexibility- no point paying $90-300/month for a dedicated server the first month if we're only getting 200 listeners on our show and 50 for the other shows but if we somehow experience a surge in listeners we don't want to be stuck at a low cap.
recommend me a reliable and cost-effective switch.
I need it to be able to do the basics and limit port speeds that aren't standard like 20Mbps or 30Mbps.
Please recommend me switches in all price ranges,
I have a Fedora Linux server and I am trying to set:
PEAR=/home/pear/bin/pear
but when I try to set it under env or in a .profile file under home it still doesn't work. I even log off and log back in and I can't just run
PEAR
I've few servers which just transffered to another data center. After switfed, my sites having the problem to be access in some area. So I thought the propagation hasn't work fully. However, yesterday my technical team found that there's actually an error at the broadcast ip(entered wrongly). So we immediately change to the correct ip.
Can we get the broadcast ip work immediately cause my sites having trouble to be displayed.
Can anyone recommend a good case/enclosure for a sata raid array? I would like to build an array using 500 GB SATA Harddrives. Will the server need much processing power and ram if I am going to use a decent hardware raid card? What card would you recommend? Are there any premade sata arrays that allow you to just pop in your own harddrives and don't cost thousands of dollars?
Also, can anyone recommend a enclosure if you had a server that had build in raid with 8 sata ports, but only two harddrive bays and wanted to use the entire 8 ports?
We're currently testing Postini after checking with Message Labs, etc and it seemed that Postini was the most highly recommended out of all of them. We shall see, as there does seem to be ALOT that get past their filters with spam level filtering set at their most sensative level.
However, what could I do for accounts with Hosting Firms. We have a couple on Pair, and while they use SA, their filters doesn't seem to be really effective at all. Users can come in over the weekend, and have 5 valid emails out of 200 junk......
BTW, has anyone used any of the spam appliances out there lately.
We tested them about 1.5 years back and none were really effective
Does anyone know of decent data centers in or around the Orlando, Florida area (possibly as far east as Tampa...) that allows 24/7 walk-in access and doesn't charge a fortune? I'm looking for a really small amount of bandwidth but I'll need around-the-clock access to the machine.
View 6 Replies View RelatedHow do you insert variables in cron jobs? e.g. the current date.
Let's say I want to back up a file and append the date to it.
0 0 * * * cp /some/file.txt /some/outputfile_[DATE].txt >/dev/null
(note that I'm not sure the cp will actually work like that, it's just an example)
Where outputfile_[DATE].txt has the date in it.
Are there variables cron can use for date, hour, minute, etc.?
I am writing a rewrite rules using mod_rewrite module. I have the same data repeating all over the rules that I would like to replace with variable and set variable once at the top of rules then use variable in the rest of rules. Then if I need to add another IP address I would just add additional IP address to the variable instead like now need to change several rules.
For example I have IP addresses that I would like to set as a variable.
Now rules are the following (simplified) in httpd.conf:
RewriteEngine Off
RewriteCond %{REMOTE_ADDR} (192.168.5.20|192.168.7.15|10.10.20.50
Which is the best and cost effective service for daily dedicated server backups?
View 7 Replies View RelatedI want to move mysql tmpdir to /ramdisk but when i am setting it, mysql errors: tmpdir is read only variable.
View 3 Replies View RelatedAnyone can tell us (from experience) a good hosting company for dedicated servers?
The most important thing for us right now is to stop a DDoS application type attack which seems to be taking place from DC++ peer to peer clients.
My server was unstable at this month sometimes fork 700 process and apache 80 access per second and that's made server very slow . very bad browsing
when i checked log files /var/log/messages found that errors
Apr 20 04:06:28 suhosin[798]: ALERT - configured request variable value length limit exceeded - dropped variable 'message' (attacker '212.107.116.238', file '/usr/local/cpanel/cgi-sys/php4')
Apr 22 00:27:05 suhosin[15442]: ALERT - configured request variable name length limit exceeded - dropped ....
PRODUCT, VERSION, MICROUPDATE, OPERATING SYSTEM, ARCHITECTURE
Parallels Plesk, 11.5, .30_build115130819.13, Debian 7.0, Intel 64bit
PROBLEM DESCRIPTION
When triggering the events 'Mail account created' or 'Mail account modified' the returned variable NEW_MAILBOX will always state TRUE regardless of the mailbox option being ticked or not ticked on the 'New mail account' option located on the 'Mail' tab of a subscription.
STEPS TO REPRODUCE
1: Log in to plesk.
2. Select 'Tools & Settings'
3: Select 'Event Manager'
4: Select 'Add New Event Handler'
5: Select Event 'Mail account updated'
6: Enter in '/usr/bin/php /root/scripts/dbmail/mail_mod.php' to the Command box.
[Code] ....
ACTUAL RESULT
[NEW_MAILBOX] => true
EXPECTED RESULT
[NEW_MAILBOX] => false
ANY ADDITIONAL INFORMATION
I submitted this with parallels support and they stated that the variable NEW_MAILBOX will always return as TRUE regardless of if the mailbox option is ticked or not. They also stated the NEW_MAILBOX variable will be true regardless of any action taken on the panel with mail accounts.
The actual documentation on this variable is very sparse
Component name/description Old component value New component value
Mailbox old_mailbox new_mailbox
We previously had some code that would trigger on this variable and would create a mail account on our mail server with or without a mailbox.
if (!strcmp(getenv('NEW_MAILBOX'),"true") && !strcmp(getenv('OLD_MAILBOX'),"false")){
//Create a mailbox on the external mail service
}
if (!strcmp(getenv('NEW_MAILBOX'),"false") && !strcmp(getenv('OLD_MAILBOX'),"true")){
//Create a mail forwarder on the external mail service
}
At some point this variable has stopped working with newer versions of plesk. Having a variable that is always true regardless of what is done when creating a mail account seems pointless and requesting a new feature that was already there in older versions doesnt seem like something that should be requested on a new feature request.
QUESTION: Is there anyway to get this functionality restored other than requesting it gets re added via [URL] ....
I run basicly run two main site.
1.Forum big one .
2.File and image sharing site.
(image sharing site generates thumbnails which produces lots of hits)
In these conditions how much difference can lighttpd can do as compared to apache for keeping my 600 MB Ram VPS host constant.