shantanuo's questions

I am trying to upgrade mysql package to 5.5 on centos. (amazon image - ami-cc5af9a5)

There is 5.1 version of mysql-libs that can not be uninstalled because postfix is dependent upon it. But I can not upgrade to 5.5 version unless I uninstall this 5.1 lib version of mysql.

# rpm -qa  | grep -i mysql
mysql-libs-5.1.61-4.el6.x86_64

# rpm -e mysql-libs-5.1.61-4.el6.x86_64
error: Failed dependencies:
    libmysqlclient.so.16()(64bit) is needed by (installed) postfix-2:2.6.6-2.2.el6_1.x86_64
    libmysqlclient.so.16(libmysqlclient_16)(64bit) is needed by (installed) postfix-2:2.6.6-2.2.el6_1.x86_64
    mysql-libs is needed by (installed) postfix-2:2.6.6-2.2.el6_1.x86_64

Can someone let me know how to uprade to MySQL 5.5 without affecting the other packages? I have downloaded rpm's because yum install / upgrade does not work since it has 5.1 version built-in.

I have 2 gz files those I need to merge into one -- that is to say I want to combine two .gz files into one such that when I extract the combined file I get a single file containing the concatenation of the original input files.

time join <(zcat r_TR2_2012-05-28-08-10-00.gz) <(zcat r_TR1_2012-05-28-08-10-00.gz)

The above statement is not working as expected. I am using 3 commands to do the needful.

gunzip r_TR2_2012-05-28-08-10-00.gz 
gunzip r_TR1_2012-05-28-08-10-00.gz 
tar -zcvf combined.tar.gz r_TR1_2012-05-28-08-10-00 r_TR2_2012-05-28-08-10-00 

and then concatenating the files together when I extract them to produce the output I want.

Is there any way to do this in 1 statement?

I am getting memory related error when I try to split a big file into 8GB files.

# split -C 8000000000 r_mini_20120503.txt  
split: memory exhausted

I am however able to create 4GB files using the same command. Is there any way to increase memory for this purpose?

The following command works on Ubuntu (GNU findutils 4.4.2) but not on Fedora (4.2.27)

time find . -maxdepth 1 -type f -executable -print

How do I find executable files using older version of find?

I want to empty out all the data on a partition. What is the quickest way to do this? It is an Amazon EBS volume if that matters.

The following is working as expected. (do not want to use nmap) I need to use nc (or any other built-in centOS) command in shell script to check the port 6379 of a remote server. I want the script to exit quickly if no response received in less than 1 second. But it seems that nc will wait for too long before quitting with exit code of 1

How do I "quickly" check if the port is listening?

# time nc -z 1.2.3.4 1234

real    0m21.001s
user    0m0.000s
sys 0m0.000s

# echo $?
1

# time nc -z 1.2.3.4 6379
Connection to 1.2.3.4 6379 port [tcp/*] succeeded!

real    0m0.272s
user    0m0.000s
sys 0m0.008s

# echo $?
0

SCP does not seem to preserve ownership stamps even if used with -p option.

scp -p /mysql/serv/data_summary.* some_server:/mysql/test/

The files are owned by mysql and I want the same ownership to be assigned on the destination server. I need to copy files as root on both servers due to some admin issues. I can not change to mysql@

The following works as expected from the command prompt. But it does not work from cron.

mysqldumpslow <(tail -1000 `mysqladmin variables \
    | grep slow_query_log_file | awk '{print $4}'`) \
    | mail -s "slow log from `hostname` sorted by time" amar`hostname`@gmail.com \
    > /root/slow_succ.txt 2> /root/slow_err.txt

I get the following error:

/bin/sh: -c: line 0: syntax error near unexpected token `('

Do certain characters need to be escaped in cron? Or is it an issue with subshell in cron?

Is it required that the server should be rebooted once I change the time? I ran into trouble when I changed the datetime and did not restart crond.

Is there any other service that may have issues after time change?

I am trying to run a script on remote server as a root user. I have this line in my shell script.

cat /home/myuser/create_user_updated.sh |   ssh myuser@$myip "sudo sh"

I get the following error.

sudo: sorry, you must have a tty to run sudo

The script "create_user_updated.sh" works ok if copy the file to remote server and execute it there locally. But it does not work from remote server.

After removing the user, does the crontab added by that user gets removed as well? I am asking this because I can see a user file called "abcuser" in the

/var/spool/cron/

When I cat the file, I can see the crons added by that user. The user no longer exist and the cron's won't work. But why does the crontab file is still there?

The following command is working as expected.

ssh [email protected] "ssh [email protected] 'cat test.txt'" > /home/shantanu/test.txt

What I need to do is to copy the entire directory instead of a single file.

Is it possible to use rync with SSH tunneling?

I use

sh -xv my_script.sh

to debug the /bin/sh script. How do I debug /bin/bash scripts?

I am getting the following error and I need to know the line that needs to edit.

# ./batch_master.sh

'@'10.10.10.10.' (using password: NO)user 'companycuser

I use the telnet command to check if MySQL port is responding.

telnet 10.10.10.24 3306

I use the ctrl character to disconnect. This is working as expected. How do I use this command in shell script?

I have a file called "nohup.out". It is consuming a lot of space and I will like to remove it.

/opt/apache-activemq-5.2.0/bin/nohup.out: 4.2G

Will it affect the application in any way? I do not want logs, only disk space.

I have a file on CentOS machine that needs to be copied to Windows NT server.

I use \10.10.10.10 to connect to the server from my windows desktop. I can copy the files using the GUI,

but I want to write a shell script on my centOS server that will periodically copy files to windows box.

The following command works from prompt but not from crontab.

 grep abc  /var/log/messages | grep "`date '+%B %d'`" | mail -s"abc log of `hostname`" s.o+`hostname`@gmail.com

I need to add it to daily cron.

How do I suppress blank emails? FOr e.g. in the following example, I will like to

some command | mail -s "tables altered on `hostname`" [email protected]

I get the following message "after" sending the message:

Null message body; hope that's ok

This is not ok. I do not want to send mail if there is no text.

How do I attach a screen? I can see a screen is running in the following output, but can not connect to it.

[root@vishnuvm6 ~]# screen -ls
There is a screen on:
        15941.pts-18.vishnuvm6  (Detached)
1 Socket in /var/run/screen/S-root.

[root@vishnuvm6 ~]# screen -x vishnuvm6
There is no screen to be attached matching vishnuvm6.

I can find my IP address using ifconfig or hostname -i command.

But how do I find my Public IP?

(I have a static public IP but I want to find it out using unix command)