Topic: Keine gezielte Wiederherstellungen aus dem Backup mehr möglich

[funky]

Can you create a bug report for the developers in perfect English?

[TerryF]

Yes, I have found the dar catalog is being corrupted for reasons as yet unknown...

There have been similar issues in the past https://bugs.contribs.org/show_bug.cgi?id=6950

Working on it

[TerryF]

Also this was an issue here as well, OS and version specific:  https://bugs.contribs.org/show_bug.cgi?id=6052#c34

[TerryF]

Before I raise a bug I want to run a few more scenarios. I will post Bug number back here when done, thanks

You are backing up to another smeserver, I am backing up to a share on a windows 10 box, cant get much more different than that, which would suggest issue is with the sme9.2

[TerryF]

OK, looks like we have found a fix, there was a patch to sme10 in late 2016 https://bugs.contribs.org/show_bug.cgi?id=9563 to fix the same issue,

It was not backported to sme9

The fix listed in that bug was not the final fix..just need to do a little more testing
and in particular try it on a big backup like my NFP 400gb one,

[TerryF]

Bug raised against sme9.2 patch will be forthcoming https://bugs.contribs.org/show_bug.cgi?id=10781

[TerryF]

see Bug 10781, rpm has been patched now verifying etc, in paerticular this, noticed it last night in early tests, make sure to read comment by Ian Wells in Bug 7035 :

"I have found that restoring a directory seems to restore all the files that are present in all the backups that are held..this needs a some more testing and maybe adding to the wiki"

[TerryF]

Small sets are fine, don't know the point at which below comes in.

On large backup sets, tested with 400+gb using selective file restore NO files or dirs are listed and the routine eventually times out..

var/log/httpd/admin_error_log: Viewed at Fri 21 Jun 2019 01:54:32 PM AEST.

[Fri Jun 21 13:40:11 2019] [warn] [client 127.0.0.1] Timeout waiting for output from CGI script /etc/e-smith/web/panels/manager/cgi-bin/backup, referer: https://192.168.11.19/server-manager/cgi-bin/backup

[Fri Jun 21 13:40:11 2019] [error] [client 127.0.0.1] (70007)The timeout specified has expired: ap_content_length_filter: apr_bucket_read() failed, referer

Work needs to be done - https://bugs.contribs.org/show_bug.cgi?id=10782

[funky]

Great, thank you for your effort. Now we are on the right path. I will wait until there is an update for the rpm package.

I have big backups over 400 GB on another machine with SME9.2. The behavior and the error is the same as on the machine with the small 224 GB backup.

[TerryF]

G'day..

With the updated rpm I am able to see all the backups including all inc on large and small backup setc, select last backup inc or full on smaller backups it will display dirs. and files etc on the test 400+gb it times out see error from logs last post, this is what we have to fix, it may NOT be fixable, I have tried increasing timeouts in httpd.conf but still errors,

I do remember in the dim mists of time a discussion about just this scenario cant find the bloody thing now it will come out eventually..

[TerryF]

If you want to play at the pointy end see the Bug 10781 on where to yum the update from or wait until it appears in updates, that could be a few days away..

Just on the smaller backup as I am hopeful this will fix the issue there..good luck

[TerryF]

Just a further note, the displaying of backup sets to select so as to list files/dirs to display in the server-manager window is working with the updated rpm.

I have been able to use dar-manager from the cli to display the list of files/dirs that the selective restore produces to be displayed in server-manager. This included a 400+gb backup, this list in most cases is quite small, also dumped the output to a small file.  So the actual cli action seems to not produce an error even with the very large backup.

Small backup sets upto at least 100gb (limit of my testing) all display a list of files/dirs in the selective restore window of server-manager.

Backup set of 400+gb (all I have to test with) times out in displaying files/dirs in the server-manager window, a smarter person than I will find the why and we can fix it. 

The "pipe/process/whatever" that sits between the perl script and server-manager (apache or whatever) breaks with the bigger backup

See Bug 8271 there were a couple of other bugs that were dupes these have been closed and consolidated in this bug

[TerryF]

Until workstation selective restore can be fixed, on the system with the very large backup set of 400+gb I have switched to using Affa and an external USB disk - https://wiki.contribs.org/Affa#Use Affa to backup to a NFS-mounted NAS or a local attached USB drive.. same USB disk as I was using with workstation backup, formatted to ntfs. Reetp is in process of updating the affa contrib, see Bug 10783, however for just a localhost backup is fine.