Error when backing up before upgrade

[c0ldf1r3]

Hello,
When following the steps to back up my installation of nagios I get to the following step:

tar cjf /root/pre-upgrade-backup/usr-local-nagios.tar.bz2 /usr/local/nagios

after executing this command the screen will just hang with the following:
[root@nagxi ~]# tar cjf /root/pre-upgrade-backup/usr-local-nagios.tar.bz2 /usr/local/nagios
tar: Removing leading `/' from member names


It never completes or comes back to the command prompt.

Any advice.

regards

[nscott]

Try this:

Code: Select all

tar cjPf /root/pre-upgrade-backup/usr-local-nagios.tar.bz2 /usr/local/nagios

[c0ldf1r3]

Scott,
I tried to run the command above but got the same results.

I execute the command and it never returns to the login prompt.

Any advice.

Regards

[nscott]

So after you include the P in the cjf you still get the

Removing leading `/' from member names

?

If you don't, and it simply doesn't return to the command prompt, that means its probably still working, compressing all of your data will take a significant amount of time.

[c0ldf1r3]

After I included the P in the cjf I DO NOT get the

Removing leading `/' from member names

However, it NEVER returns the command prompt and I let it run overnight.

any advice.

regards
John

[nscott]

Well, depending on how many hosts/services you're checking and how long you've been checking them for, the /usr/local/nagios directory is going to be absolutely massive. I would make sure you have enough hard disk space. To get an idea of how big that will be:

du -hc /usr/local/nagios | grep total

Thats how much stuff its going to have to pack. If this is virtualized you could take a snapshot of its current state, otherwise the backup will just have to runs its course. You could just backup Nagios XI, which could keep your config files

http://assets.nagios.com/downloads/nagi ... ing_XI.pdf

But this will not keep your performance data.

[c0ldf1r3]

Scott,
I get the same results when running du -hc /usr/local/nagios | grep total

It never return a result after running all weekend.

As fas as disk space:
Filesystem 1K-blocks Used Available Use% Mounted on
/dev/mapper/VolGroup00-LogVol00
123760492 43279476 74092896 37% /
/dev/sda1 101086 28620 67247 30% /boot
tmpfs 6601080 0 6601080 0% /dev/shm

Getting very frustrated with this whole install.

[nscott]

Running 'du -hc /usr/local/nagios | grep total' did not return a value? That shouldn't take much processing power at all and should return immediately.

I suggest running the tar job with the verbose flag. And since its taking so long I also suggest turning the compression down a notch. bz2 compression takes a lot of CPU horspower to run. You could change that the just .gz. Also turning on the verbose flag so we get this:

tar cvf <backup location> /usr/local/nagios

If you are not a seeing a stream of filenames then there is something wrong, if it sits on a particular file for longer than around 10 minutes then its a similar situation.

[agriffin]

If 'du' is also hanging then this honestly does not sound like a problem with Nagios. It could just be that that folder is very large, as nscott suggested. Since we can't run du to see the size of the directory, we'll have to estimate. Approximately how many hosts & services are you monitoring?

Also, it could be related to inode usage. If you could post the output of the command 'df -i' we could check that theory.

[c0ldf1r3]

the output of fd -i is as follows:

[root@nagxi ~]# df -i
Filesystem Inodes IUsed IFree IUse% Mounted on
/dev/mapper/VolGroup00-LogVol00
31948800 18697770 13251030 59% /
/dev/sda1 26104 42 26062 1% /boot
tmpfs 1650270 1 1650269 1% /dev/shm

This install of Nagios has approx 350 hosts with 35 groups. I would not think it would considered large by any stretch.

Any advice is greatly appreciated as I really can't do a reinstall at this point.

Thanks