Jump to content

Why does Dantz make the use of SELECTORS so difficult?


churchtown

Recommended Posts

New upgrade/update... same old trouble... happens every time... me trying to figure out EXACTLY the right sort of stuff to put into which ever SELECTOR clause to stop all the errors appearing in the logs;~/

 

 

 

It's not rocket science Dantz! A particular file isn't reachable, readable, available, whatever... why can't the core code just give me an option... "Do you want to avoid this file in future?" instead of me having to second guess the inner workings of the core code and work out PRECISELY what type and form of SELECTOR filter to input to persuade Retrospect Professional v.6.5 to avoid that particular troublespot in future runs.

 

 

 

Please tell me the design reasons, I assume there are more than one, why I the long-suffering user should be FORCED to try and work out whether the latest file or directory, with which Retrospect is having its latest difficulty, is an OS directory, user file, Martian path, or this that the other etc...?

 

 

 

Dantz, you are my file specialists! Why is it that your programme has always ceded this apparently necessary level of technical knowledge/understanding back to the punter!?

 

 

 

For the record my online (picture) galleries are -now- causing the usual (upgrade/update) grief in the Retrospect logs. So, if I may be so bold, I'm passing the operational buck back to your designers. My v6 has always backed up my galleries OK but now v6.5 seems to be baulking at doing the same job as least as well. The error lines that I am now required to arrange for Retrospect to workaround are:

 

 

 

File "\\SYSTA\bb\html\gallery\setup\": can't read security information, error -1017 (insufficient permissions)

 

 

 

File "\\SYSTA\ic\html\gallery\setup\": can't read security information, error -1017 (insufficient permissions)

 

 

 

These are directories on my server whose permissions are deliberately closed as a security measure. It is not necesseary to have them backed up ordinarily and it is not necessary to have to be advised of Retrospect's inability to access them on each backup run either.

 

 

 

Your precise advice would be appreciated as to precisely what I have to input into precisely which type of selector clause in order to quash these somewhat unnecessary and unwanted error [sic] report lines.

 

 

 

best wishes, Robert

Link to comment
Share on other sites

Ah! And I thought it was all metaphorical dust and spiders in this otherwise very well presented forum. A fellow human!:-)

 

SELECTORS is easy enough to 'use and/or misuse'. The tricky bit is forward-guessing quite what the original issue that baulked Retrospect and quite how the SELECTORS processing can/does affect future accesses. All technical. All fairly unnecessary for the average punter. Not even particularly necessary for sysadmins (like me) who would also like to live a life occasionally.

 

best wishes, Robert

Link to comment
Share on other sites

keep in mind that often you just can't do anything about those errors. Suppose you have some files with closed permissions and you even exclude from backup because you just don't care about them. Still you get those errors because retrospect, even if it's not backing these files, still try to access them when it creates the snapshot.

Don't ask why retrospect included all files in the snapshot (even those that have been excluded): it's beyond me.

Link to comment
Share on other sites

lbertacco, I do respect your viewpoint. From my own POV I intend to work a little at attempting to get this addressed. Accepting the routine presentation of errors, false positives or otherwise, is dangerously close to The Slippery Slope To Ruin... ;~/

 

A routine report should feed no errors from a properly run unattended backup.

 

When the system-generated email arrives, bells ring and whistles blow for immediate action if that message should be parsed 'with errors'.

 

A properly run unattended backup report message is parsed 'without errors' and is properly and automatically filed away unread, unsullied by mere human intervention.

 

Continual false alarms, unfortunately, will always tend to breed an attitude based on contempt, indolence, apathy and a general lack of attention... need I say more? I said... Dantz! Need I say any more? Wake up please, it's so very quiet in here;~)

 

best wishes, Robert

Link to comment
Share on other sites

Quote:


File "\\SYSTA\bb\html\gallery\setup\": can't read security information, error -1017 (insufficient permissions)

 

File "\\SYSTA\ic\html\gallery\setup\": can't read security information, error -1017 (insufficient permissions)

 

 


 

The backup of NTFS permissions are a totally different process that takes place after files are copied. Even if you exclude a file from the backup, Retrospect will still try to copy the NTFS permissions.

 

The Microsoft API used to copy permissions does not know what a Retrospect selector is, so it attempts to copy all permissions that exist on the source volume.

 

See my other post in this forum section for selector tips.

Link to comment
Share on other sites

Now I am confused;~/ There are no NTFS permissions anywhere on \\SYSTA\. That thing has only ext3 journal format SCSI drives. It's a Linux server (my gateway/www/email server). Nothing Micro$oft-flavoured is ever allowed anywhere near it.

 

If it helps any, I use NTFS on my workstation \\SYSTD\ because the TV capture card makes 4GB+ files, otherwise I would cheerfully reformat everything on my W2KPRO workstations back to FAT32!

 

best wishes, Robert

Link to comment
Share on other sites

Interesting, my guess is that Retrospect is still copying the linux permissions as part of a separate process the same way it handles NTFS permissions.

 

Because the permissions are not copied along with the file itself, the engine used to copy the permissions does not know you have excluded the files.

 

On Mac, Windows and linux permissions are an "all or nothing" event, we are not able exclude them in a selector.

Link to comment
Share on other sites

I have a little experience in M$ and some in Linux, but I have never even seen a Mac (in the flesh as it were) let alone used one or have any experience of its OS.

 

Will now experiment with the settings for the automatic script. Earlier I ticked some security settings in the forlorn hope (mmmm;~/) that they might resolve one of my 'other' difficulties with RS6.5 (it keeps forgetting my volumes shares' logins during an unattended script).

 

best wishes, Robert

Link to comment
Share on other sites

I turned off the two Windows' security settings in the script settings on the options page... these are ones referring to NTFS and/or server permissions.

 

 

 

I have also unticked the Open Files option settings because strangely enough I do not have XP but RS6.5 doesn't seem able to accept this straightforward fact and otherwise produces yet another false positive error report line:~/ ( Subject of a new thread...? )

 

 

 

Attended running of the script:

 

Litany of errors - particular errors STILL includes the two silly permissions errors under discussion

 

 

 

Unattended running of the script:

 

STILL largely impossible - unattended mode causes RS6.5 to forget the shared volume login passwords (subject of an, as yet, unresponsive Dantz forum thread). The shared volume login information --IS-- remembered by RS6.5 after a manual exit/entry of the programme but forgotten during, or because of, the attempt to run the unattended backup [sic] script:~|

 

 

 

Dantz "Do you want just the 5min discussion or the full hour's worth...?" to mis-paraphrase an otherwise infamous Monty Python comedy script. We're talking scripts. The latter made me laugh. The RS6.5 scripts are just making me cross.

 

 

 

And which thread do I continue? Is it that the SELECTORS are causing the trouble...? Or is it that the SCRIPTS are causing trouble...? Or is it that volatility of the shared volume login password information is causing trouble...? Or am I just going around in ever-decreasing circles? ...it used to be so easy running backups...

 

 

 

best wishes, Robert

Link to comment
Share on other sites

None of my above worked. I still cannot make any unattended network backups work - RS6.5 just will not remember the network login password. Dantz assistance would be appreciated. This is driving me nuts. I am getting further and further behind in my work, simply because of all this unexpected work just attempting to get Retrospect to perform a backup successfully.

 

best wishes, Robert

Link to comment
Share on other sites

churchtown, I perfectly agree with you and I hate getting these errors reported even when they shouldn't. What I was saying is that while you keep retrospect to create the snapshot, it will try to access ALL files from the backed up volume, even files excluded from backup.

I consider this to be a retrospect design flaw.

 

Also even if your filesystem is ext3 and not NTFS, retrospect is ofcourse reading the CIFS/SMB ACL permissions that samba reports according to the CIFS protocol.

Link to comment
Share on other sites

Hello

 

This thread is a bit hard to follow so I'd like to review if I may:

 

Do you get error 1017 errors on the entire volume or just certain files?

 

What login information are you using in the security preferences for Retrospect?

 

Have you manually edited the launcher service in any way?

 

Thanks

Nate

Link to comment
Share on other sites

Quote:

This thread is a bit hard to follow so I'd like to review if I may:

 


I am sorry about that Nate. After many days have passed the situation changes so much. I'm more used to real time or near real time responses from manufacturers or their technical support forums. I have elsewhere admitted and apologised for my earlier mistaken perception of this peer-to-peer forum. Yes, a review would be most welcome and I am glad to be able to assist the sole, active, representative of the host organisation.

 

 

Quote:

Do you get error 1017 errors on the entire volume or just certain files?

 


For a specific answer to the subject of this thread... sub-directories (two special ones, one in each photo gallery on two websites of mine).

 

For a less specific answer, elsewhere... volumes.

 

The tense is now DID (past tense) as I now have most of these -1017 errors fixed in RS6.5 except for the two special sub-directories that cause -1017 errors (see below for further information).

 

 

Quote:

What login information are you using in the security preferences for Retrospect?

 


I cannot give a simple answer to that question. The security preferences section of RS6.5 would not accept my username/password, it kept aborting citing an inability to assign privileges. Sorry Nate but I forget the exact phraseology. I have been shown so many negative, seemingly irrational, error phrases from RS6.5 that they are jumbling up in my memory, particularly so long after the events have passed.

 

 

Quote:

Have you manually edited the launcher service in any way?

 


Yes, in an effort to imitate what I used to have to do in RS6.0 but I found that this just seemed to stop the launcher service triggering RS6.5 (to then do a half-baked job). I read elsewhere on the Dantz site that this area, I assume it's the same area, of RS had been fixed in 6.5 and so undid what I had earlier attempted.

 

Later on I uninstalled RS6.5 and reinstalled in an effort to gee RS along and actually getting to perform. That wasn't very effective either. It wasn't a 'clean' install, much Registry information was retained and reused. It didn't 'repair' what I later found to be the prime difficulty (RS6.5 unilaterally does not recognise that I am indeed THE administrator and would not therefore do anything remotely sensible). It also overwrote (downgraded) the RDU49 update I had already applied, which confused me in yet another issue (my NEC burner not being natively and properly used by RS6.5 to burn 8x DVD-r discs) until I realised what had happened and overwrote what RS6.5 reinstallation had overwritten with the driver 4.9 updater.

 

Technology don't you just love it... I'm a photographer, it's a much simpler way of getting technology to do things;~)

 

I am not sure what other points of this thread you would like to review but here are a few pro-active comments:

 

* SELECTORS will not stop RS6.5 inappropriately accessing forbidden website sub-directories on my gateway/website/email server and so feeding further inappropriate errors into the report on every backup run. Apparently it's a Micro$oft thing and Dantz have to toe the M$ doctrinal line in this respect. This is not something I favour nor can I find it in myself to condone. There ARE ways around M$ and I intend to adopt any and/or all of these. The most important being the increasing use of Open Source / Linux.

 

I am now searching for a more effective and much simpler way of backing up these things as it now appears that Retrospect (6.5) is continuing to prove problematic and whose issues are seemingly intractible in this particular iteration (Retrospect Professional 6.5 non-server)

 

I have been otherwise advised that the (RS) server Client is 'the way to go' but the price that Dantz continues to put on the head of anything remotely connected with Linux exceeds what my small business can possibly justify. If this attitude does not change I cannot see a continuation of the existing Dantz customerbase (home/SoHo users).

 

best wishes, Robert

Link to comment
Share on other sites

Quote:

churchtown, I perfectly agree with you and I hate getting these errors reported even when they shouldn't. What I was saying is that while you keep retrospect to create the snapshot, it will try to access ALL files from the backed up volume, even files excluded from backup.

 

I consider this to be a retrospect design flaw.

 


 

lbertacco, we are in mutual agreement:-) I will add that I think that the RS design flaw seems to be the result of an unnecessary adherence to toe the M$ line and an apparent unwillingness to do something about it on their own.

 

 

 

Quote:

Also even if your filesystem is ext3 and not NTFS, retrospect is ofcourse reading the CIFS/SMB ACL permissions that samba reports according to the CIFS protocol.

 


 

Errr... I'll have to take your word on that ;~) I'm a relatively simple chap on those matters.

 

 

 

best wishes, Robert

Link to comment
Share on other sites

Hello,

 

Selectors allow you to remove files from your backup however they do not remove them from the volume snapshot. This is by design as Retrospect must take an image (snapshot) of the entire volume regardless of what files were actually backed up. This process is a cornerstone of Progressive backup.

 

In other words if the error 1017 is happening during snapshot creation there isn't much you can do to avoid it.You could turn of snapshots (not advised) or upgrade to Windows XP which will allow Retrospect to backup open files.

 

Is your web server running while you run your backup?

 

Retrospect saves very little information in the Registry. Settings are stored in c:\documents and settings\all users\application data\retrospect config65.dat and .bak

 

The only time I have seen problems with Retrospect not logging in as administrator properly was when there were other permissions problems on the machine. Retrospect uses a standard windows logon so I doubt this is a bug in Retrospect.

 

Thanks

Nate

Link to comment
Share on other sites

Nate

 

No, I won't be turning off snapshots. We are agreed that that would be inadvisable.

 

As for the suggested upgrade [sic] to Windows XP, that is out of the question. The only true upgrade path permissable here is to move all the workstations off W2KPRO-SP4 and into Linux.

 

The problem area for RS6.5 is not one of 'backing up open files' (whether this is supported in the unacceptable move into WinXP or not) as they are NOT open files at all but protected/secured sub-directories. They are quintessentially NOT open but completely and utterly closed - absolutely unavailable.

 

For anyone curious about this thing, my galleries use Gallery (Open Source) v1.4.2 software...

http://gallery.menalto.com

...their operating procedure involves both chmod and .htaccess which set appropriate access permissions on the server.

 

Yes, my 'web server' is running and continues to run when the unattended backup starts itself. And, yes I agree, taking a gallery offline to run the backup would enable the previously secured directories (with respect to access permissions etc) and thus avoid causing RS6.5 to flag an error. It would have to be done manually and done during working hours. I am unsure of my competence at combining win/linux programming to achieve this level of unattended automation.

 

Yes, RS6.5 must have left its user-orientated information elsewhere than the Registry. My uninstall/reinstallation exercise retained just about all, if not all, of my previous settings and so forth. I mentioned it only in passing.

 

* * * * *

 

Nate, I have to take complete issue with you concerning RS6.5's asinine Judgment that my normal, reasonable and actual Windows OS login is NOT that of the administrator. RS6.5 has been utterly unique in that decision.

 

To be perfectly honest I have NEVER, before now, heard of any Windows programme that has effectively refused to accept the administrator's login as being the one and only true administrator. Doesn't make sense. If I am the su (in Unix/Linux terms the superuser) then I am the su. Period. Please elaborate, I'd be really interested to listen to your take on this RS6.5 issue;~)

 

I am the system administrator (of the entire site) and login thus whether RS6.5 deems me the administrator or not.

 

If RS6.5 cannot accept that my login is that of the system administrator then I am forced to insist that that *IS* a bug in the Retrospect v6.5 Professional upgrade download software - as supplied to me from Dantz.

 

To corroborate this please review my 2½hr trouble report file with Dantz EuroSupport, presumably they have it logged. It should outline the issue and subsequent resolution - that of bringing up a new WinOS 'user' with admin group permissions and switching RS6.5 over to running under the alternative new user's permissions.

 

best wishes, Robert

Link to comment
Share on other sites

Hi

 

This is a side note but...

An uninstall of Retrospect does not uninstall the Retrospect settings file that I mentioned in my last post. You have to delete that manually. As to why that is the case I have no clue. Regardless, that is where your settings are stored, not the Registry.

 

If you never deleted the settings files it may be why the problems continued. A corrupt retrospect config file would retain the login information that was not working properly.

 

I'm fairly confident that the admin login issue is not a bug in a Retrospect. I have seen it work too consistently to think it is a flaw. I agree it is a problem but I think there is an explaination somewhere.

 

Here is my theory:

In a past install of Retrospect 6.0 you had manually altered the launcher service. You then installed Retrospect 6.5 which automatically imported your old Retrospect 6.0 settings.

 

Retrospect 6.5 eliminated the need to manually configure the launcher service because it offers the windows log on information. Retrospect 6.5 now passes that information to the launcher service so you don't have to.

 

I think that your old launcher service and settings files were causing a conflict so the log in feature was not functioning properly. To remedy the issue I would have suggested a full uninstall of Retrospect, Reboot, removing the Retrospect configs folder and trying again with another clean install.

 

I Imagine you tried that with David.

 

Here is the other reason I think this isn't a bug:

There are some rare cases where Retrospect or Retrospect client will not be able to see a local volume on a computer. In those cases we invariably find that permissions are not set up correctly on the volume in question. Either something was corrupt in windows or someone/some application changed them.

 

In your case there could be a number of things happening. The hostname of the computer may not be being recognized properly, the administrator account may be corrupt in some small way, it is hard to say.

 

Were you able to get around this problem by creating a new admin user on the machine and setting Retrospect to use that account?

 

Thanks

Nate

Link to comment
Share on other sites

Nate

 

An uninstall should do just so, if only to obviate a subsequent occurrence of programme corruption after (possibly in desperation) going through an uninstall/reinstallation cycle. My opinion.

 

This admin login issue has only afflicted and affected this flawed RS6.5 upgrade from RS6.0... so, from my user point of view, I call that a bug in Retrospect but I cannot appreciate why it is not from your manufacturer's POV. Certainly glad you see it as a problem to be fixed though!

 

I feel let down by this implementation of Retrospect and I believe I have made this abundantly clear. The published spec/design of RS6.5 fixed the issue where previously (WinOS?) RS6.0 users were instructed by Dantz/Retrospect to manually patch the RS6.0 Launcher in the WinOS Admin Services area. I thought it reasonable (still do in fact) that the RS6.5 fix included the automatic undoing of the previous and required workaround/patch. If not, why was I not routinely advised, during the upgrade installation process, to manually undo the previously required patch? RS6.5 has let me down, caused a 4day hole in my productivity, still makes me feel sore (obviously) and Dantz have done nothing to ameliorate my intense annoyance over their flawed RS6.5 upgrade installation process.

 

David and I did not have recourse to have to uninstall (again) and then manually go around wiping all traces of RS from my workstation. As I have already indicated, the action we took to remedy the flawed upgrade process of RS6.5 was to invoke a new user attributed with admin group rights and force RS6.5 to adopt that new user. To answer more in context a later point you raised, that is NOT the same as 'creating a new admin user'. I have not changed THE administrator at all.

 

I have no knowledge of those rare case to which you refer. Or of what you refer to as being the a local volume. The flawed RS6.6 upgrade was put into the booting C: drive partition. I would call that a local volume but, there again, I'm not network trained in any way, shape, or form. I just use my intranet and expect/hope/pray it continues to work the way I expect it should. And this is pretty much what it does, I do have to say:-)

 

best wishes, Robert

Link to comment
Share on other sites

Archived

This topic is now archived and is closed to further replies.

×
×
  • Create New...