Request-URI Too Long

Plural

HI Marcin,

I'm getting this error when I try to restore some Jobs, not all. I have no problems restoring from bconsole. This happen using the Restore wizard or the restore button in the job detail.

Best regards,
Rui

gani

Hi Rui,

Thanks for reporting this problem.

In this wizard are sent some GET type requests to the API. It looks that in your environment for some jobs to restore the GET size limit is exhausted (HTTP Error 414). It can be caused e.g. by selecting a job that has many related jobids, for example many incremental jobs.

I would propose to do a try and next time if this problem happens, could you please run this bconsole command with the selected to restore jobid?

.bvfs_get_jobids jobid=YOUR_JOBID_TO_RESTORE

For example:

.bvfs_get_jobids jobid=49647

This command returns all particular jobids required to prepare restore with YOUR_JOBID_TO_RESTORE. I guess returned list will be long. We will see.

To avoid receiving this HTTP Error 414, in your Apache web server I would propose to check this parameter:

https://httpd.apache.org/docs/2.4/mod/core.html#limitrequestline

If you did not define it in your web server config, then the default value is 8190 bytes that is quite long. After increasing it, this error should gone.

If it will be this case I will add to the Bacularis documentation information about this error and how to work with it.

Please let us know about your tests.

Thanks.

Best regards,
Marcin Haba (gani)

Plural

Hi Marcin,

gani This command returns all particular jobids required to prepare restore with YOUR_JOBID_TO_RESTORE. I guess returned list will be long. We will see.

Yes, you are right. This failing jobs have a lot of incrementar jobs. I will run a Virtual Full backup in this jobs. For this I'm want to use the Bacularis wisard, but in step 3 (Where) the storage list is empty.

gani If you did not define it in your web server config, then the default value is 8190 bytes that is quite long. After increasing it, this error should gone.

That sounds like a workaround, I thing the virtual full backup it's the way to go.

Best regards,
Rui

gani

Hi Rui,

Thanks for letting us know about this problem with empty storage to virtual full control.

I confirm that this is a bug in 5.2.0. I am trying to fix it but finding right fix is not trivial.

If you want to use this wizard without waiting on fix, temporary you can use this oneliner:

diff --git a/Web/Layouts/Wizard.tpl b/Web/Layouts/Wizard.tpl
index dee1ddf..a2d2d94 100644
--- a/Web/Layouts/Wizard.tpl
+++ b/Web/Layouts/Wizard.tpl
@@ -34,7 +34,6 @@
                        <!-- Top container -->
                                <com:TContentPlaceHolder ID="Wizard" />
                </com:TForm>
-               <com:Bacularis.Web.Portlets.ConstantPicker />
 <script>
        Formatters.set_formatters();
        $(function() {

Plural That sounds like a workaround, I thing the virtual full backup it's the way to go.

Great. Good luck with this virtual full.

Best regards,
Marcin Haba (gani)

Plural

Hi Marcin,

Thank you for this.

Best regards,
Rui

gani

Hi Rui,

We have fixed this storage list in the new virtual full job wizard.

https://github.com/bacularis/bacularis-web/commit/5e2875ebcf4a391cce9296c88814f0554feb897b

This fix will be available in the next release.

Thanks.

Best regards,
Marcin Haba (gani)

Plural

Thank you Marcin.

Best regards,
Rui

Plural

Hi Marcin,

I'm back to report my findings and solutions about this issue.

The Virtual Full backup, aka consolidation, solves the problem.
Also, some of our Jobs does not copy any files because there are no changes in the source. Because of that, they are a lot job records in the catalog that are good for backup confirmation but not relevant in terms of future search and restore of files. I created a python script thats find and deletes this records from the catalog.

For incremental jobs than run many times during the day and keep track of small modified files, like our motion fx and graphics departments, the Virtual Full backup is the best way to solve the issue.

For incremental jobs than run some times during the day and keep track of large new files, like our recorded video files, deleting the zero files jobs it's the way to go.

Best regards,
Rui

gani

Hi Rui,

Many thanks for sharing your solution for this situation. This can be useful for other users that will encounter this problem to use the virtual full.

I think this is a good idea with this script. I use something a bit similar for some selected jobs finished with errors to remove them. This works periodically in the Admin job.

Best regards,
Marcin Haba (gani)