I have somewhere between 8 and 12 servers, spread amongst different webhosts, all running CentOS 5.3. Everything on the servers has been compiled from source: > EngineX, PHP, MySQL, Munin, etc
I compile from source because I usually run ./configure to modify the source before installing. What would be the easiest way to keep the server software updated? I am considering creating custom packages from a server which I would keep up to date manually. Then use Capistrano or Puppet to install those packages on all of my servers.
I have installed CentOS 5.3 with Xen on a PowerEdge 2650 machine with 6GB of RAM. As usual, PAE was already enabled so I did not have any problem with utilising all of the memory. However after upgrading to the latest release of the kernel (2.6.18-128.7.1.el5xen), memory available decreased to 4GBs. Then I switched back to the old kernel and 6GB was there. Then finally have switched back to the new kernel and 6GB is again there. So now PAE seems to be enabled but what concerns me is the inconsistent behaviour. (Also I am not sure the reason was the upgraded kernel.)
I have a copy of the DVD Iso for centos 5.3. I downloaded the updated packages to the Centos directory and then ran the repomanage perl script to remove the old files from the directory. I then ran the createrepo and the new iso image with the script code below.
I am using VMWare to test the build, so I have the cd pointing to the iso image. I get the CentOS to start up find and dandy asking the questions for the interactive boot. It gets thru the stage of checking dependencies and then when it starts to copy down the image to the "harddrive" that is when the problem occurs.
One of the updated files is file-4.17-15.el5_3.1.i386.rpm (file-4.17-15.el5.i386.rpm was removed using repomanage), but the loader is looking for the removed file. I've looked thru any dependencies, but nothing specific for the removed file, all are asking for /usr/bin/file with no specific version numbers. I have run a rpm -test on all the rpms, but haven't been able to look thru that to see if there is a specific request for the version.
I did try this, but it just moved on to the next file. I did not replace the file version, but then it found another problem that was the same as this, the updated file is in the repo, but it is requesting the old version. I looked thru the fileslist and others to see if maybe that was the problem, but they were updated to the new versions.
I came to a "weird" situation when I saw in yum log files that several centOS systems have been automatically updated via yum. I checked the following: 1. chkconfig for init services that could upgrade the system: only yum-updatesd 2. In the config file of yum-updatesd.conf the do_update function is set to no. 3. If an updated made manually by a user ..no. Where else should I look or how this update happened?
Recently I updated to centos 5.5 , every thing is fine , but the mkinitrd can't updated
When try: yum clean all yum update
The output : Loading mirror speeds from cached hostfile * addons: mirror.cogentco.com * base: pubmirrors.reflected.net * extras: mirror.highspeedweb.net * updates: mirror.cogentco.com Excluding Packages in global exclude list Finished Setting up Update Process .....
Error unpacking rpm package mkinitrd-184.108.40.206-61.el5_5.1.i386 error: unpacking of archive failed on file /lib/bdevid/ata.so: cpio: rename Failed: mkinitrd.i386 0:220.127.116.11-61.el5_5.1 Complete!
Currently our Production Server version is Fedora8. I know its very old version, i was newly joined as server admin for this company.. my first task need to Upgrade Server with all updated packages and patches..Without production time down..because we have nearly 400 clients accessing our server.
1. Is it possible to do Without Production loss??
2. before upgrade what are the things i need to do??
3. is there any possibles the working function not working in new upgrade packages??
I haven't updated my centos box for almost about a year, my theory is if it ain't broke don't fix it.These last two weeks I've been replacing hardware, a motherboard, memory, processor, and a hard-disk. Everything worked fine with the current 2.6.18-128.1.10, but I decided that I should upgrade to the latest along with all the new packages. But in doing so I would need to recompile the driver module for the raid controllers, rr1740. So I recompiled them using the 2.6.18-164.1 build while I was in kernel 128. When I was done, I restarted the machine and when it booted to the new 164 kernel, it came back very long error messages but here's what I can remember off the top of my head.
What is the best place to get an updated nVidia video driver for CentOS 5.5? Is there a CentOS package available that will update it? Or is it best to download it directly from nVidia? Also, I do not have an internet connection on the machine, so it will have to be a manual download and installation.
i am integrated ADS with squid and its working fine. In squid server end i have used "net ads password" to update new password for the user and it successfully updated. Issue is update not doing immediately it takes long time to update the password , even i restarted the smb and winbind services. I want to updated the password immediately in the ads server. is this possible ?
want to sync 2 folders, one on a desktop and the other on my server. My objective is to keep the desktop folder always updated with the content of the server folder. If I get this working, I can do the same with the rest of my desktop and laptop users. When online they can run a script with rsync and update data. Is it possible to get 2 way sync?
I have configured master and slave Bind servers. Everything works fine. But whenever I add a new zone entry at master server it is not getting updated at slave server in logs I see this error: client 192.168.1.1#43428: view external: received notify for zone 'yourdomainname.com': not authoritative
At master server I do not see any error or warning message. This error clearly indicates that named.conf file does not have zone entry in it or domain name is wrong. While checking the named.conf file I see that the zone entry has not been updated at slave server. If I update it manually and reload named on slave then zone files (db files) are getting created without any issue and any modification at master server for the zone records are also getting updated. My concern is why zone record is not getting appended at slave server in named.conf file.
Is there anything I am missing in the configuration. I am pasting the steps which I have followed to configure my master and slave server: Configure Bind as master and slave server Install Bind on your server yum install bind OR sudu apt-get install bind9 Generate RNDC Key using the command rndc-confgen -a -k rndc-key it will stored in /etc/rndc-key file Master Server IP 192.168.0.1 Slave Server IP 192.168.1.1 Master Server Configuration options .....
we're about to migrate a set of workstations (ubuntu 10.04 LTS) to a new kerberos/LDAP setup. Basically, this requires the installation of some required deb packages and to copy some new .conf files over the original ones.We made a deb package having these "features":requires the needed other packages as dependenciesbacks up original conf filescopies the new conf files to the right places (i.e. /etc/krb5.conf,/etc/ldap.conf)The problem is: apt-get complains because the deb is "touching" files owned by other packages (kerberos, ldap, etc.). Therefore, the only way to skip this check is either to force apt-get to proceed or using the "replaces" directive in the deb control file, specifying the clashing packages. omething like this:
I have a problem with compiling of mod_ruby-1.3.0 After a succesfull configure i get a error in the make, it say "make: *** [apachelib.o] error 1" . Here below you can find the results of de configuring and the make.I was following the roles in Howtoforge " The perfect server - Centos 5.3" everything goes perfect till mod_ruby-1.3.0
./configure.rb --with-apr-includes=/usr/include/apr-1 checking for a BSD compatible install... /usr/bin/install -c checking whether we are using gcc... yes
I recently moved over user from an old box running postfix(v 2.0.16) over to rhel 6 running postfix (v mail_version = 2.6.6). ive tried to make sure all the files are of correct permissions and that the main.cf file is configured corectly. However there is something wrong as when i run postfix: service postfix server i get no error but when checking the status:service postfix status i get: master dead but pid file exists Looking into /var/log/mailog i find this line being the issue:
During the install of my Centos server 5.3 i cam in trouble after setting the Static IP address, Gateway and DNS i cannot connect to the internet my own network is visable. When i start ifconfig my network is running.
I wish a easy server with CentOS. I need only: ssh, ftp for backup , sftp for over the web, apache2 php5 mysql webserver, easy install printserver for a laserprinter with a com-port, and in the future a samba server. make a choice which version i need with CentOS. I am a beginner at all.I only know things about ubuntu server. I like to know the differences.
i have been trying to complete the following project1) Configure a FTP server where we can upload and download files.........2) server must run at 9 pm & stop at 9 am automatically ............although the first task was easy ,i have no idea how to accomplish the 2nd task(not to mention I'm a new user)