Month: January 2011

Unmount your ext2 filesystem, then issue:
# tune2fs -j /dev/hda1
Where /dev/hda1 is the ext2 FS you want to convert.
then edit your fstab and change /dev/hda1 entry to be ext3
Easy!
210/433

to know how many cpus do you have there:
# cat /proc/cpuinfo
to know memory info
# cat /proc/meminfo
254/433

ok, just hang with me while I learn NB.
some jobs were stalled in a machine, so we went to the media server and checked if the jobs were pending
and the tape drives were not in “TSD”:
bash-3.00# /usr/openv/volmgr/bin/vmoprcmd
PENDING REQUESTS
<NONE>
DRIVE STATUS
Drv Type Control User Label RVSN EVSN Ready Wr.Enbl. ReqId
0 dlt TSD – No – –
1 dlt TSD root Yes EBI219 EBI219 Yes Yes 1
ADDITIONAL DRIVE STATUS
Drv DriveName Multihost Assigned Comment
0 QUANTUMDLT70000 No –
1 QUANTUMDLT70001 No ebitdb1
In the “pending” section there were jobs pending on drv 1 and 0, and the tape drives were unavailable, so we
reset them:
# vmoprcmd -deny 0
# vmoprcmd -deny 1
# vmoprcmd -up 0
# vmoprcmd -up 1
We assumed the drives were recently fed with tapes, so after this jobs were running ok.
To see the active jobs…
#bpdbjobs -report |grep -i act
To see the queued jobs…
#bpdbjobs -report | grep -i que
To kill a job
#bpdbjobs -kill <job#>
To see the job summary…
#/usr/openv/netbackup/bin/bpdbjobs -summary
#bpps -a
To run an inventory on a tape library…
#bpinvent 0
259/433
Check last 24 hour’s errors
#bperror -U -backstat -hoursago 24
See available tapes
# available_media
Or go to bpadm – media mgmt – special actions – Inventory a Robot and Compare with Volume Configuration
To see if a tape is in the correct pool
#vmquery -m <mediaid>
To change a tape of pool
# vmchange -h ntxldb2 -p 1 -m NTXD53
(Where 1 is NetBackup and 2 is systems)
To see the pools
#vmpool -listall
A monitoring loop:
# while true;do clear; vmoprcmd; echo “active”; bpdbjobs -report |grep -i act;echo “queued”; bpdbjobs -report
|grep -i que;bpdbjobs -summary; bperror -U -backstat -hoursago 24| tail -15l; sleep 5; clear; done
Run bpadm for interactive actions
# bpadm
document
too
260/433

If you’re trying to save your /etc/fstab and you get:
“fstab” E212: Can’t open file for writing
And you know you can write any other file in “/”, then it’s not a readonly FS.
You need to check the file’s attributes.
[root@myserver etc]# lsattr /etc/fstab
—-i——– /etc/fstab
if it has a “i” then it’s write protected, so change it
[root@myserver etc]# chattr -i /etc/fstab
[root@myserver etc]# lsattr /etc/fstab
————- /etc/fstab
Then you can edit the file
75/433

We got the output from “sar -ruq” in a file, we stored in sar2.txt, and it looks like this:

11:44:51 633773 22954346
4 1 3 93
0.0 0 0.0 0
11:44:56 633773 22954288
2 2 1 94
0.0 0 0.0 0
11:45:01 632206 22786120
13 18 13 56
9.0 20 0.0 0
11:45:06 625509 22462685
53 47 0 0
5.0 99 0.0 0

So we want to plot it in excel, we need one line per sample, much like this:

11:45:06 625509 22462685 53 47 0 0 5.0 99 0.0 0
11:45:11 628749 22770298 59 41 0 0 4.8 100 0.0 0
11:45:16 630432 22733837 55 45 0 0 6.8 100 0.0 0

So we wrote a small gawk expression to do it:

gawk 'ORS=NR%3?" ":"n"' sar2.txt

So simple, we’re using the modulus function here to do the job. Coming to a blog near you!

I wrote a script that shows you in a condensed way if you’re in network card speed trouble:
# ——————— server app213
bond0= eth1 eth2
bond0 : : : : : 10.20.9.38
eth0 : 100Mb/s : Half : on : yes : 192.168.5.119
eth1 : 100Mb/s : Half : on : yes :
eth2 : 100Mb/s : Half : on : yes :
eth3 : Unknown! (65535) : Unknown! (255) : on : no :
Here it goes:
#!/bin/bash
DIR=/tmp/.dir
mkdir $DIR > /dev/null 2>&1
HB=`ifconfig -a | grep -i bond | wc -l | sed ‘s/ //g’`
if [ “$HB” -ne “0” ];then
for i in `ifconfig -a | cut -c0-10 | egrep -i ‘bond’| grep -v ‘:’`;do
echo -n “$i=”
for i in `cat /etc/sysconfig/network/ifcfg-$i | grep “BONDING_SLAVE” | awk -F”=” ‘{ print $2 }’ | awk -F”‘” ‘{
print $2 }’`;do
echo -n ” $i”
done
echo “”
done
fi
for i in `ifconfig -a | cut -c0-10 | egrep -i ‘bond|eth’| grep -v ‘:’`;do
SP=`ethtool $i | grep “Speed:” | awk -F: ‘{ print $2 }’`
DU=`ethtool $i | grep “Duplex:” | awk -F: ‘{ print $2 }’`
AU=`ethtool $i |grep “Auto-negotiation:” | awk -F: ‘{ print $2 }’`
LI=`ethtool $i | grep “Link detected: ” | awk -F: ‘{ print $2 }’`
(for w in `ifconfig $i`;do
echo “$w”
done) | grep “addr:” | awk -F”:” ‘{ print $2 }’ > $DIR/addr
echo “$i : $SP : $DU : $AU : $LI : `cat $DIR/addr` ”
done
set network card speed by default in redhat.
If your interface is eth0, in the etc/sysconfig/network-scripts folder, vi the ifcfg-eth0 script. add the parameters
ETHTOOL_OPTS=”speed 100 duplex full autoneg on”, and you’re set
Next time the server reboots you have the right speed set
215/433
216/433

Yeah, we all go and put a script on /etc/rc2.d or /etc/rc3.d
But to be clean on Solaris 8, 9 and 10 you can add your static routes at /etc/gateways like this:
net 100.100.100.0 gateway 192.1.243.1 metric 1

for Solaris 10 11/06, we can use the “-p” parameter

/usr/sbin/route -p add 175.168.0.0 192.168.2.1

these routes are stored in /etc/inet/static_routes
have fun

$ w | egrep -iv ‘USER|average’ |awk ‘{ print $1 }’ |while read u;do echo ” :)” | write $u; done
10/433

To check if the system can see the HBAs
# luxadm qlgc
To verify HBAs are connected
# luxadm -e port
199/433

if, when creating partitions with fdisk or cfdisk you get this message on exit:
WARNING: Re-reading the partition table failed with error 16: Device or resource busy.
The kernel still uses the old table.
The new table will be used at the next reboot.
Syncing disks.
You don’t need to reboot the box, just issue “partprobe”