When removing graphs by command line, regex is not properly validated when empty

[bmfmancini]

Hey Guys

This may be or bug or by design but if you run
php remove_graphs.php --graph-regex= --list
you will get a list of all of the devices in your install
Is there supposed to be an error displayed if the regex is null?

This is going to work great for me because I am in the middle of changing our data profile so I need to delete all of my graphs but I could see this being a danger if someone were to issue the force option and forget to add a regex value .....maybe...possibly ??

Here is the output from my test

[root@li1303-178 cli]# php remove_graphs.php --graph-regex= --list
There are 4 Graphs to Remove. Use the --force option to remove these Graphs.
ID GraphName
17 Local Linux Machine - Processes
18 Local Linux Machine - Load Average
19 Local Linux Machine - Logged in Users
20 Local Linux Machine - Memory Usage

[root@li1303-178 cli]# php remove_graphs.php --graph-regex= --force
There are 4 Graphs to Remove. Data Sources will be removed is possible.
[root@li1303-178 cli]#

[bmfmancini]

Oh BTW if it is a bug could we add keep it as a feature say something like
php remove_graphs.php --all
with a confirmation that your about to do something bad(Potentially)
I know that its not something you would do every day but it would come in handy if you need to

[netniV]

That would work currently because the regex is considered not to have been set, rather than being set to something that will never match. If you had used ' ' after the =, that would have only listed those with a space.

Should that be considered a bug, I think so. If you are setting the regex, you should have a regex.

[bmfmancini]

ok cool
can we add the --all flag ??

[cigamit]

Why don't you just change the profile, and repopulate the poller cache, move the RRDfiles to a backup location, wait for the new RRDfiles to be created, and then use the rrd_splice.php tool to stitch the old and the new together?

[bmfmancini]

I would but there are 35k graphs and tha kfulky we are ok for a reset! Time saved!

…

On Thu., Jan. 30, 2020, 7:41 p.m. Jimmy Conner, ***@***.***> wrote: Why don't you just change the profile, and repopulate the poller cache, move the RRDfiles to a backup location, wait for the new RRDfiles to be created, and then use the rrd_splice.php tool to stitch the old and the new together? — You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub <#3224?email_source=notifications&email_token=ADGEXTHGBW57RR7B54ZJRHDRANXUVA5CNFSM4KNLY4N2YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOEKNCVZI#issuecomment-580528869>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/ADGEXTE2VJSH2TBYDKJ3GJ3RANXUVANCNFSM4KNLY4NQ> .

[bmfmancini]

Oh and my goal is to change the averaging and the retention not the poll time so the full rra would need to be recreated just curious I guess the rra split wouldn't do so well in that case? On Thu., Jan. 30, 2020, 8:05 p.m. Sean Mancini, <sean@seanmancini.com> wrote:

…

I would but there are 35k graphs and tha kfulky we are ok for a reset! Time saved! On Thu., Jan. 30, 2020, 7:41 p.m. Jimmy Conner, ***@***.***> wrote: > Why don't you just change the profile, and repopulate the poller cache, > move the RRDfiles to a backup location, wait for the new RRDfiles to be > created, and then use the rrd_splice.php tool to stitch the old and the new > together? > > — > You are receiving this because you authored the thread. > Reply to this email directly, view it on GitHub > <#3224?email_source=notifications&email_token=ADGEXTHGBW57RR7B54ZJRHDRANXUVA5CNFSM4KNLY4N2YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOEKNCVZI#issuecomment-580528869>, > or unsubscribe > <https://github.com/notifications/unsubscribe-auth/ADGEXTE2VJSH2TBYDKJ3GJ3RANXUVANCNFSM4KNLY4NQ> > . >

[cigamit]

If you are doing something like adding rows to rra's, the splicer will still work. Just make a backup copy of the file before you add rows, and then splice the old into the new. You will see no gaps in the data for areas of rows that were added.

The splicing can be time consuming, so it's best to do it in parallel with as many threads as possible.