Jump to content

DavidHertzberg

Members
  • Content count

    1,038
  • Joined

  • Last visited

  • Days Won

    38

Posts posted by DavidHertzberg


  1. Indi-Tech,

    My little brain is thoroughly confused by "the client (primary server)" in the third paragraph of your immediately-preceding post, and by "hosts a share that is used by backup server because the backup server also hosts a share that requires backup" in the final paragraph.  Please rewrite that post using the terms "LAN data server" and "Retrospect backup server" to distinguish the separate functions of these machines/drives on your LAN.  You'll need that rewrite, if for no purpose other than to communicate clearly with Retrospect Technical Support, which I suggest you do—ASAP to get your free personalized T.S. if possible.

     


  2. Fredspon,

    As this post in a 2012 thread shows (click the blue-underlined link to be taken to it, then click the Back arrow in your web browser to be taken back here), "it's not a bug, it's a feature!" :)  There's a zero-byte file named "Backup Media" in the "Retrospect" folder on the "backup disks" Member(s) of your Backup Set that's designed to activate the feature.

    The reason Dantz Development Corp. (the predecessor of Retrospect Inc.) put in this feature years ago is to prevent administrators from "holding it wrong" (that's a Macintosh joke, referring to a Steve Jobs on-stage remark about a previous version of the iPhone), as this post (responding to the one linked to in the preceding paragraph of this post) says.

    Quote

    The correct method for what you are trying to achieve is to create a second backup set on the offsite media, then use "Transfer" (not "Duplicate") operations to copy files to it from the onsite backup set.

    If you've scheduled a Transfer operation (not "program"; you'd better be more precise about terminology if you want us to help you :rolleyes:) after each Backup operation, then use Transfer Snapshots—here's the Tutorial that you should watch after the one linked-to in the next sentence.  If not, use Transfer Backup Sets—here's the Tutorial that you should watch before the one linked-to in the preceding sentence.  For the scripted operations, see pages 206-217 of the Retrospect Windows 15 User's Guide; for the Immediate operations, see pages 141-152.


  3. As many of you know, I follow the 20th Century practice of backing up onto one of 3 portable HDDs formatted with HFS+, one of which I rotate each week—after 7 preceding days of use—into a bank safe deposit box.  Each Saturday, when I rotate the next of my portable HDDs back into use, I Recycle its associated Media Set—first with a "sacrificial" script and then with a "real" script that backs up 3 drives and a Favorite folder.   My MacBook Pro's SSD—formatted with APFS since 2 April—is the first one backed up in the "real" Recycle script on Saturdays, and the only one backed up in the Saturday "sacrificial" script.  After that SSD (the only drive on my MacBook Pro) was reformatted with APFS on 2 April 2019, I got:

    • 6 April 2019 "Media Set White"—which had been backed up on 16 March starting with HFS+ SSD: OK on both "sacrificial" and "real" scripts
    • 13 April 2019 "Media Set Blue"—which had been backed up on 23 March starting with HFS+ SSD: "sacrificial" script was stopped after many  "!Error reading snapshot record at ...", Rebuild from portable HDD, OK on "real script
    • 20 April 2019 "Media Set Red"—which had been backed up on 5 April starting with HFS+ SSD : OK on both "sacrificial" and "real" scripts
    • 27 April 2019 "Media Set White": "sacrificial" script was stopped 16 minutes after "!Error reading snapshot record at 61653107172u", Rebuild from portable HDD, OK on "real script
    • 4 May 2019 "Media Set Blue": "sacrificial" script failed with "!Can't reset Backup Set Media Set Blue, error -1,102 (drive missing/unavailable)", "real" script showed No Media Action—stopped, Rebuild from portable HDD, OK on "real" script
    • 11 May 2019 "Media Set Red": "sacrificial" script failed with "!Can't reset Backup Set Media Set Red, error -1,102 (drive missing/unavailable)",  Rebuild from deleted Media Set Red on portable HDD, OK on "real" script

    I also run "sacrificial" scripts, which use the No Files rule, before the No Media Action "real" scripts I run to backup my MBP the other 6 days of the week.  There have been no problems with these No Media Action "sacrificial" or "real" scripts.

    I wasn't having any of the above  Recycle problems with my portable HDDs up through 30 March, while the SSD on my MBP was still formatted with HFS+ .

    P.S.: Slightly cut down so the part through the itemized list could fit in the Problem Statement, I have turned the contents of this into Support Case #67777.


  4. Indi-Tech,

    Apologies for not realizing until now what your problem probably is. 

    First, for background, you should watch this Retrospect Mac Tutorial video again—even if you have watched it already.  Note that, at minute 1:47, the last thing the head of RTS does is add "client" computers—and he does it by first specifying the pre-defined name for the NIC and then specifying the method of scanning for "client" computers.  I strongly urge you to specify Use Subnet Broadcast, and I hope that "is added manually with the 'service' subnet" means that's what you did.

    But the key requirement IME is that the Client on each "client" not yet have a "backup server" address stored in a "secret field" within it at that point—which is the situation assumed in the Tutorial, meaning it hasn't yet been imprinted with the address of the "backup server" with which it is supposed to communicate.  That imprinting is done by the "backup server" when you first Add a "client"; therefore (as told to me 4 years ago by Andy :)—now departed from RTS) you must Uninstall and then Install the Retrospect Client on each "client" computer.  I have done this in the past by downloading the latest Retrospect Client .zip from the website, rather than messing with trying to accomplish this from the Console.  But before running the Retrospect Client Installer.pkg for the "client" you must have first Removed that "client" computer from Sources via the Console.  Only after running the Retrospect Client Installer.pkg on the "client" computer can you then re-Add the "client" into Sources from the Console—via Use Subnet Broadcast—and thus imprint the "secret field".  Up through Retrospect Mac 15, the "secret field" where the Client stores forever (I'm not sure whether Refresh can change it) the identity of its imprinted "backup server" used to be displayed—uneditable—in the Advanced tab of the Client; version 16.0.2.101 displays 0.0.0.0 instead for what is now called the Private Backup Server field—which may merely be a byproduct of Engineering hijacking the associated dialog control for use in editing the newly-added Public Backup Server field when needed for Remote Backup.  The unchangeable-after-imprinting function of the "secret field" is informed speculation, BTW, since I don't have access to the Retrospect source code.

    After you've done this for each "client" computer (the sequence I recommend is:  [1] do all "client" Removes from the Console, [2] do the Client Uninstall and Install on each "client" computer, and [3] do all "client" Adds from the Console), you'll have to re-add the "clients" to each Backup/Archive/Proactive script that uses them.  After doing all "client" adds for a script by selecting the script in Scripts and then check-marking the boxes for all desired "clients"' in the panel shown by clicking the script's Sources tab, you should Save the script and then click its Summary tab.  If the "clients" don't appear in the backing-up sequence you want, you can—with some GUI kludginess—drag their names in the Summary list pane into the desired sequence (thanks derek500 :)); if you've dragged, do another Save of the script.

    If you don't believe me, read this final post in the Forums thread I linked to up-thread.  Nigel Smith says what I suggested didn't quite work for the problem he was having, but that it was because his problem was dealing with a Fortigate setup—with "client" computers connecting at random to only one NIC at a time—rather than the multiple NICs you evidently have on your version of the Mac Pro (3,1).

    P.S.: Changed 2nd substantive paragraph to get rid of male-chauvinist-like language, and clarify that paragraph—I hope.

     


  5. Indi-Tech,

    Since you're using Retrospect Mac version 16, look in that User's Guide at page 64 for "Network Interfaces" and pages 78-81 for "Advanced Networking".

    I have no knowledge of VLANs; I just now took a look at the appropriate Wikipedia article.  However you might find an answer in this relatively-recent Forums post and its successors by an administrator who is much more knowledgeable than I am about such matters.

    I have also noticed that the version 16 Mac Client's Advanced tab has a new Public Backup Server entry capability.  As is unfortunately par for the course in recent versions of Retrospect, the "What's New" chapter of the UG is pure marketing drivel, and doesn't discuss that capability.  However you might try clicking the lock icon and entering your "Client"'s password to make changes, then clicking the Edit button and entering the address of the "backup server".  I must caution you that Your Mileage May Vary, and see the P.P.S. below.

    Finally, how did you generate the screenshot in your most-recent preceding post—the one you posted as if it were a copy of another post?  Was it with the Retrospect "backup server", or with another application?

    P.S.: I suggest you phone Retrospect Technical Support at (888) 376-1078 or (925) 476-1030‬; they're on extension  3.  If you bought Retrospect Mac 16 within the past 30 days, I believe they'll give you personalized help for free.

    P.P.S.:  The new Public Backup Server entry capability is described in this section of a Knowledge Base article, and is intended for use in Remote Backup (introduced in version 15, so nothing remains in the UG).  But IMHO that doesn't mean you couldn't pervert it to do what you want.  You could speed the perversion by simply installing a server.txt file on each "client" machine, as described in "Installation Process" in the KB article.  However using this capability would for now require converting your scripts to Proactive, so you'd do better to phone RTS instead.


  6. 7 hours ago, x509 said:

    Yes, I forgot to mention grooming a dataset with selectors.  I'm having a real issue with Retrospect ignoring a some folders that were explicitly excluded, and this grooming has allowed me to recover about 1 TB of wasted space on my 6 TB (5.45 TB in Windows) backup drive.  See this thread, particularly the last few posts.

     

     

     

    x509,

    Isn't it rather ironic how the decision of some Brussels bureaucrats (with which I, as a computer user, generally agree) caused Retrospect Inc. to add a valuable new feature?  Of course there was a temporary price for Retrospect Windows administrators, as mentioned in the fourth paragraph of this post in another thread.

     

     


  7. Indi-Tech,

    You don't say which version of Retrospect Mac you are using, so I'll refer you to the Retrospect Mac 15 User's Guide.  "Network Interfaces" on page 60 sounds applicable to your problem. 

    For further information, see "Advanced Networking" on pages 74-77.  I suggest that —at a minimum—you add your "client" machines with Subnet Broadcast on your "service subnet".  If you feel more energetic, you can Add Source Directly—for which you will have to assign each "client" machine a fixed address on your router using its MAC address.  That's what I do on my home LAN (with "backup server" and "client" now running Retrospect Mac 16), for which Multicast stopped working a couple of years ago for mysterious hardware-related reasons.  But I have only one subnet, so Add Source Directly may not solve your problem —whereas Subnet Broadcast probably will.


  8. Assuming, x509, that "I want to confirm that that approach works" means "I want someone else to confirm that that approach works" rather than "I want to confirm that that approach has worked for me", I've done the following—using Retrospect Mac 16.0.2.101:

    Every morning except Saturdays at 3:05 a.m. I schedule a No Media Action (Retrospect-Mac-speak for Normal) backup of my MacBook Pro "client"'s only drive, but 5 minutes before that "real" script is scheduled I still schedule—out of habit although it's no longer necessary— a "sacrificial" script that does the same backup for that same "client" with the No Files Rule (Selector).  x509 doesn't pay me enough money ;) to stay up until 3 a.m., so last night before I went to bed I re-scheduled this morning's "sacrificial" script for 3:04 a.m.. My BPH awakened me around 3:40 a.m., upon which I booted my MBP followed by my Mac Pro "backup server".  The "sacrificial" script" immediately started running, while the Activities panel (the Retrospect Mac equivalent of merged panels from all the Activity Monitor tabs except History)  on my Console showed a yellow warning icon—which when clicked said "Waiting for Media Set White" (which I would have named Backup Set White if I had been running Retrospect Windows)—scheduled before the "real" script.  As soon as the "sacrificial"  script had finished running in about 4 minutes, the warning icon disappeared and the "real" script—with the All Files Except Cache Files Rule (Selector)—began to run.

    This is the behavior I've had since I started scheduling "sacrificial" scripts with Retrospect Mac 12 in February 2017.  So, barring any unlikely scheduling difference between application variants, I predict "that that approach works" for Retrospect Windows too. :)

     


  9. 2 hours ago, x509 said:

    David,

    I meant the ability to transfer backup files from one dataset to another backup dataset, in the interest of eliminating redundant file backups and simplifying the number of old datasets to store and maintain.

    x509

    But Transfer Snapshots and Transfer Backup Sets have been in Retrospect Windows since 7.7, although there were some bug fixes as late as Retrospect Windows 11.5.  Am I missing something about your wish for Retrospect Windows 15, x509?


  10. 12 hours ago, x509 said:

    Nigel,

    I tested out your suggestion in Windows 10 Explorer.  The trailing backslash is not necessary for Windows to understand that I am referring to a directory and not to a filename.  Same is when I do a "Save As" with different applications such as Microsoft Word or Adobe Lightroom.

    It's a damn selector bug, methinks.

    x509

    Surely, x509, I don't need to tell you why and how to file a Support Case for a bug.


  11. x509,

    Sorry "life has been happening" for you, :( but welcome back to the Forums.

    By "the transfer datasets features of V15", do you mean

    Quote

    Configuration Import/Export: Transfer options for transfer snapshot script now included (#5580)

    or something else in a previous release?

    Since the beginning of April 2019 I've been using Retrospect Mac 16.0.2.101.  One minor problem I've had that may affect everybody upgrading to version 16 concerns scanning Media/Backup Sets with a Recycle script.  I do this twice each Saturday as I Recycle the Set whose sole Member is on one of my three portable HDD drives (I'm very 20th century; I swap one of the drives every Friday in my bank safe deposit box, because until last October I didn't have the upload speed for Cloud backup—whatever that is ;)).  Early Saturday morning I run a Recycle backup of all my 6 drives, but 5 minutes before that is scheduled I still run—out of habit although it's no longer necessary— a "sacrificial" script that does the same backup for my first "client" with the No Files Rule/Selector.  The last 3 Saturdays I've found the scanning phase on the "sacrificial" script running very slowly, because of errors such as

    Quote
    
    !Error reading snapshot record at 61653107172u: wanted 11 MB, had 10.9 MB

    Once I notice this is happening I pause all scheduled runs and stop the "sacrificial" script, then  do a Rebuild of the Media Set being Recycled, and have no further problem for the remainder of the week with that Media Set—even when doing the same Recycle with the afterward-un-paused "real" Saturday backup.  I didn't have the same problem when I upgraded to Retrospect Mac 15.6.1.105 in January 2019, and the problem may be peculiar to Retrospect Mac 16.  But be warned that, if any of you downgrade from Retrospect 16.0 to Retrospect 15, you may have to do a Rebuild of your Media/Backup Sets from the existing Members.

    BTW, x509,  I noticed that in your post directly above this one you made the same mistake I made in this post in the thread where lhlo accused me of "condescension".  You initially thought the first letter of "lhlo" was an upper-case 'I' (the letter whose name  sounds like "eye") instead of a lower-case 'L'.  Maybe it would have been more polite if, instead of saying lhlo had picked a cutesy-poo "handle", I had merely suggested he/she either take a course in Human Factors or repeat the (Canadian equivalent of?) third grade.


  12. DovidBenAvraham has been faced with a new problem by this revised Knowledge Base article, which now says:

    Quote

    Known Issues

    • Retrospect cannot perform a bootable restore or bootable duplicate for an APFS destination volume. We are in touch with Apple to resolve this issue.

    • Instant Scan is not compatible with APFS.

    The second issue sounds as if it is a permanent result of design differences between Apple's old HFS+ filesystem and its new APFS filesystem (with one result that Retrospect Mac administrators have already discussed in this Macintosh 9+ thread).   The first issue, however,  sounds as if it's something for which the Retrospect Inc. engineers may eventually develop a workaround.  Therefore DBA has chosen to put mentions of the second issue into the Wikipedia articles here and here, but to treat the first issue as a "temporary bug" that need not be mentioned in the articles.


  13. What the head of Retrospect Tech told me in this Agent Response turns out to have been officially announced the day before in courageous ;) fashion—which I guessed is what might happen.  This revised section of an existing Knowledge Base article now says:

    Quote

    Known Issues

    • Retrospect cannot perform a bootable restore or bootable duplicate for an APFS destination volume. We are in touch with Apple to resolve this issue.

    • Instant Scan is not compatible with APFS.

    Enjoy the weekend. :rolleyes:


  14. Nigel Smith,

    I didn't want to say before, because I thought it would be distracting.  My Late 2016 MacBook Pro came with macOS 10.12 Sierra in December 2017; I bought it open-box precisely because I thought High Sierra was too new.  A couple of months ago I upgraded it to High Sierra under the direction of an Apple Support person (he suggested it to fix a problem with several mini-apps in Dashboard); the upgrade file he told me to download did not re-format the SSD as APFS (presumably he had the "technical knowledge").  The backup I was originally trying to Restore, made on Saturday 30 March, was of the MBP's Macintosh HD formatted with HFS+.  My original attempted Restore (stupidly) over the LAN—which stopped because of  "insufficient permissions"—was made to the HFS+-formatted  MBP running High Sierra.  My external HDD, onto which I later Restored using Retrospect 15.6.1.105 running under Sierra on my Mac Pro "backup server", had probably previously contained a Restore of my old MBP—which IIRC was running OS X 10.10 Yosemite when its logic board died in December 2017.

    Here's something that may blow your mind.  If I boot my MBP with the Option key held down, it goes directly to my ordinary High Sierra desktop.  That may be because there is no other drive cabled to the MBP.  I'll haul the external HDD from the bedroom into the study, cable it to my MBP with the USB-C-to-Firewire adapter string, and try that again; but not now, because I was up past 3 a.m. writing the posts above and copying the last reply as an Additional Note to my Support Case—so I'm going back to bed.  Late this afternoon I'll wander over to Mike's Tech Shop and try to find out if they did anything tricky to my MBP's firmware on Tuesday 2 April; the Ethernet adapter they sold me in December 2017 is a Moshi one sold under Apple auspices.


  15. The third Agent Reply to my Support Case #67374—and IMHO it really is explosive—was: 

    Quote

    Agent Response:

    >Mike's Tech Shop install macOS 10.13 High Sierra—which forcibly formatted my MacBook Pro's SSD as APFS

    The issue is that the operating system and data is being restored to an APFS volume. AFPS has specific system requirements that make a restore of the full OS very complex and not supported by Retrospect at this time. We have been working with Apple on a long term solution for this limitation.

    You may be able to do the following:

    Restore the original system's user directory only onto a fresh Mac OS system installed on the APFS disk. Then create a new user account exactly matching the name of the user's directory, let the permissions set on the directory, and then log in to that user.

    Please let us know if you have any additional questions,

    My Additional Note to that will be approximately [after replacing any links to posts in this thread with Support Case date-time-stamps]:

    Quote

    Dear Robin,

    While you are working with Apple on a long-term solution,  why not try a fairly simple short-term solution based on what worked for me on Tuesday 2 April?  In essence, it is to execute versions of the procedures in the four bulleted paragraphs in my OP in this thread:

    • Install the appropriate version of macOS—10.13 High Sierra or later—on the boot SSD of the machine being restored, which will format the SSD with APFS. 
    • Obtain unrestricted use of a spinning rust HDD that has (1) enough capacity to contain a complete restore of the target machine,  and (2) a port or port-plus-adapter(s) that allow connection of the HDD to the target machine via FireWire or Thunderbolt.  Connect the HDD to your "backup server", erase the HDD and format it—with Ignore Ownership un-checked—with HFS+.  Then do a complete Restore of the the contents of the SSD of the machine being restored.  It doesn't matter if the HDD ends up un-bootable, or if there are some errors during the Restore.
    • Connect the HDD to the target machine with FireWire or Thunderbolt, boot the target machine, and run Migration Assistant (which may call itself Install Assistant) to copy as much of the contents of the HDD as possible onto the target machine's SSD.
    • Boot the target machine from its SSD.  If there's fancy software—such as DisplayLink—that wasn't successfully copied by Migration Assistant, Google installers for that software and re-install it on the target machine.

     


  16. The second Agent Reply to my Support Case #67374  was:

    Quote

    Agent Response:

    As noted in the original reply, we specifically said HFS+.  

    If the internal Macintosh HD is an APFS volume, the process is very different and requires a bunch of terminal commands.  APFS volumes have a PreBoot volume and if that volume is lost or missing, the steps are totally different from an APFS volume that still has the PreBoot volume.  We have support cases open with Apple to try and find the best solution for users.  The easiest approach with an APFS volume is to reinstall the Mac OS and probably just restore your Users folder instead of trying to restore an entire OS. 

    Please let us know if you have any additional questions,

    My first Additional Note response to that was [as edited to replace Support Case date-time-stamps with links to posts in this thread]:

    Quote
    
    Let's first discuss what happened to me on 30 March, as described in [the first two 
    paragraphs of my OP].  It may not have been sufficiently clear from the last sentence of 
    that [second paragraph], but up through noon on Sunday 31 March my MacBook Pro's SSD had 
    always been formatted with HFS+, and its Recycle backup on Saturday morning was backed up 
    all the way to 1-Backup Set Red with HFS+.  So APFS isn't a factor then.
    
    What it looks like instead is System Integrity Protection kicking in for an Apple-installed
    application in Macintosh HD/Library.  According to this section of an Ars Technica article,
    if I had known to do so (a different error message number and text in Retrospect would have
     been informative ) I could have _first_temporarily_ disabled SIP by booting into the 
    Recovery partition and invoking csrutil from Terminal.
    
    The Ars Technica Mac Forums post 
    https://arstechnica.com/civis/viewtopic.php?p=37119277#p37119277 says "Time Machine, for 
    example, would not have had an issue doing the restore. I know, I've used it to do a 
    restore on multiple occasions."  So let's convert _this_post_ into an Enhancement Request: 
    Add to Retrospect the capability of doing a _temporary_ disabling of SIP for the duration 
    of a full-disk Restore of a Mac source disk.  If you folks feel that needs an extra 
    check-mark box, then put one in.

    My second Additional Note response to that was [as edited to replace Support Case date-time-stamps with links to posts in this thread]:

    Quote

    What you are suggesting in the only substantial paragraph of [Agent Reply at the top of this post] is IME far too pessimistic.

    Take another look at the [third paragraph of my OP], along with [all the bulleted paragraphs of my OP] and [the final paragraph after them in my OP]. What I did was to have Mike's Tech Shop install macOS 10.13 High Sierra—which forcibly formatted my MacBook Pro's SSD as APFS, and then to use Migration_Assistant to copy over nearly all the the necessary other files from the non-bootable Restore I had done onto my external HDD. (The only problem was getting my external 27-inch Apple LED Cinema Display working again, which I solved Tuesday night 2 April in an hour or so by finding out how to re-download and re-enable the omitted DisplayLink software.)

    So IMHO the question really is: Why did my external HDD _repeatedly_ Restore as non-bootable, with around 90 cases of "[*] MapError: unknown Mac error 22" followed by one ""!Trouble writing folder "/Volumes/Macintosh HD OS X+/.HFS+ Private Directory Data", error -1,101 (file/directory not found)"—which from the timestamp immediately ended the Restore? As I said in [my response in this up-thread post], I tried doing the same Restore from a cumulative Media Set made from my MBP after its SSD had been formatted with APFS, with almost precisely the same errors. (My external HDD remains formatted with HFS+, which makes it easy for me to FireWire-cable it to my Mac Pro "backup server" and use Disk Utility to re-initialize the HDD). If Engineering can give me a version of Retrospect Mac 16 with enhanced logging options, I'll be glad to re-test the Restore again.

     


  17. The first Agent Reply to my Support Case #67374, whose Problem Statement was a slightly-edited version of my OP in this thread, was:

    Quote
    
    Agent Response: 
    
    If the disk you backed up was running an HFS+ File system, the best option is to install 
    the Mac OS into an external disk. Boot from that disk and while booted from that disk, 
    install Retrospect.  Then perform a complete Restore of the entire disk back to the 
    internal drive, while booted from the external volume.  
    
    Please let us know if you have any additional questions.

    My Additional Note response to that was [as edited to replace Support Case date-time-stamps with links to posts in this thread]:

    Quote
    
    Thanks, Robin, but a test I performed this evening makes me _question_ whether what you 
    suggested—which is very complicated—would work.
    
    I basically repeated what I described in the [second bulleted paragraph of my OP].  Because 
    I have since Recycled the 1-Media Set Red drive I used on 2 April, it now contains (along
    with other Recycle backups) a Recycle backup plus 3 No Media Action backups of my MacBook 
    Pro—which is now formatted with APFS.  The external Firewire HDD remains formatted with 
    HFS+.  The only difference in the results was a slightly smaller number (in the high 80s) 
    of "[*] MapError: unknown Mac error 22".  Given the results recounted in the last sentence 
    of [my OP], I haven't even bothered to try booting my MBP from my external Firewire HDD—it
    almost certainly wouldn't work.  So why would doing a complete Restore of my MBP while 
    booted from that external HDD work, just because the external HDD had previously been 
    prepared with a macOS install and a complete install of the Retrospect Engine and Console?
    
    I think the procedure described in the [third bulleted paragraph of my OP] is less trouble.
    It does not assume that a complete Restore of the of the entire MBP SSD would produce a 
    bootable SSD, just because the Restore is being done under control of the external SSD 
    rather than under the control of my Mac Pro "backup server".  And what is your 
    explanation of why the procedure described in the first paragraph of [this post] 
    repeatedly didn't produce a bootable HDD on the external Firewire HDD—either when the 
    restore was of a HFS+ backup or of an APFS backup?
    
    What you have suggested is essentially the "Restoring a Mac using an Emergency Tools disk" 
    procedure on pages 149-150 of the Retrospect Mac 16 User's Guide.  What I've successfully
    done is a variant on "Using FireWire Target Disk Mode" on pages 146-149, not using Target
    Disk Mode because I didn't have a USB-C-to-Firewire adapter.

     


  18. Nigel Smith,

    Command-Option-Shift-R directly following  powering-on my MacBook Pro boots into the same macOS 10.13 High Sierra environment it normally boots into, including all my Documents and such non-Apple applications as Firefox and Retrospect Client.  That contrasts with doing the same thing on my "backup server" Mac Pro, where Command-Option-Shift-R directly following  powering it on immediately brings up a dialog box offering me Disk Utility etc..  So it's not that I'm "holding it wrong"; it appears to be the fact that my MBP's SSD was formatted with a single APFS partition by Mike's Tech Shop, while the SSD my Mac Pro normally boots from was formatted with HFS+ and populated with macOS 10.12 Sierra after the SSD was inserted into one of the Mac Pro's drive trays by little old me.

    IM very HO Mike's Tech Shop would have had to use the Terminal options of Disk Utility and possibly csrutil on my MBP to have "managed to get that back to HFS+ first".  That probably goes beyond what they're willing to do for US$75, and the employee who did it may not know how to do that at all.

    I received a third Agent Response from the head of Retrospect Tech Support this morning.  The news in that is IMHO so explosive that I need to put it into context when posting it here, and I don't have time to do that before my physical therapy appointment.  My macOS version was 10.13 High Sierra under HFS+ as of Saturday night 30 March, but IMHO the real problem was with the incompleteness of the Retrospect 15 Restore onto my external HFS+-formatted HDD.  So have a good night's sleep, and the appropriate info will all be here for you tomorrow.

    P.S.: SIP does indeed protect Apple-provided applications in /Library; read far enough down this section of the Ars Technica article describing it to look at the list generated by

    cat /System/Library/Sandbox/rootless.conf

     


  19. 3 hours ago, Nigel Smith said:

    What OS was the MBPro on? Asking because that sounds more like Apple's new(-ish) security sandboxing kicking in, rather than an UNIX permissions error.

    If you fancy trying again, reboot with Command-Option-Shift-R held down -- that should install the OS version the Mac shipped with (unless that's the "Apple database" you mention).

    I'll confess, we long-ago decided to only back up user data while reinstalling the OS, apps, etc from scratch. Always seems like a good opportunity to clear out the cruft that accumulates over time. But for bare-metal restores in the past I've always connected the Mac-to-be-restored to another Mac (either client or the RS server) in Target Disk mode, though I guess lack of adapter prevented you from doing that initially.

    I've a spare laptop sitting on the bench right now. Let me know the client OS and I'll set up a test (though it'll be with Server v13) using the "Live Restore" instructions here.

    You're right, Nigel Smith, what I described in my first paragraph in this thread's OP looks like System Integrity Protection kicking in for an Apple-installed application in Macintosh HD/Library.  According to this section of the Ars Technica article, if I had known to do so (a different error message number and text in Retrospect would have been informative :rolleyes:) I could have temporarily disabled SIP by booting into the Recovery partition and invoking csrutil from Terminal.

    As described in my second and third paragraphs, after messing around with Jenna's help on Sunday 31 March I couldn't boot my MacBook Pro into the Recovery partition.  Moreover, in trying to do so I think I would have run into the same "Apple database" problem Mike's Tech Shop ran into if I wanted to revert to macOS 10.12 Sierra—which wouldn't have forcibly formatted the MBP's SSD in APFS as High Sierra did.

    I thank you for your offer to set up a test, but I don't think that will be advisable or necessary.  The "advisable" comes from my awareness that the text in the Retrospect Mac User's Guide—other than the "What's New" chapter—hasn't been updated in around 5 years (the same is true for the Retrospect Windows UG), so I rather doubt that the "Live Restore" instructions still work for recent versions of macOS.  The "necessary" comes from my MBP having been restored to satisfactory functioning per the last bulleted  item in this thread's OP.

    I've converted this thread's OP into Support Case #67374, and have already received two responses from the head of Retrospect Tech Support. :o   I'm inclined to let him deal with the "bare metal" Restore problems or turn them over to the engineers. :)


  20. Saturday night 30 March I hastily decided that I needed to do a full-SSD restore of my Late 2016 MacBook Pro.  Because I didn't have a USB-C to-Firewire adapter, I started to do the restore from Saturday morning's Recycle backup over my LAN.  This was stupid, because the restore would have wiped out my Retrospect 15.6.1.105 Client, but it didn't get that far.  After "Finished deleting 2,385 unnecessary files and 198 unnecessary folders on destination", it bombed with  "!Trouble writing folder "Macintosh HD/Library/", error -1,017 (insufficient permissions)".  How does one avoid that with Retrospect, since I'm told Time Machine avoids it?

    Meanwhile I was left with a MBP SSD with no usable macOS on it. I next attempted to download and install Sierra on the SSD, expecting to then be able to solve the permissions problem and rerun the backup application's Restore of my complete High Sierra disk. However I couldn't find how to do that download of the installer, and it was by then after Apple Support's 11 p.m. EDT closing time. I should explain here that, for reasons explained in this KB article and this Forums post, I'd prefer for now to keep all 6 drives on my 3 Macs using HFS+ rather than APFS.

    So I phoned Apple Support around noon Sunday. I spent a couple of hours talking to the senior Support technician Jenna. I first downloaded a High Sierra installer, which left me with a single APFS partition. Disk Utility wouldn't let me erase that partition, or delete it because it was the first one created. I worked on the problem for another hour while Jenna took a four-hour break; I ended up with an SSD which sat forever with 7 minutes to go when I booted into the Recovery Partition, and stopped booting after a couple of minutes—showing a big white circle crossed by a right-slanting diagonal line—when I did a conventional boot without Command-R—etc..

    Here's what happened after I took my MacBook Pro in to Mike's Tech Shop on Monday 1 April:

    • Despite my (open-box 15-inch Late 2016) MBP having come with macOS 10.12 Sierra installed, Apple has some database that says it can't run under anything earlier than macOS 10.13 High Sierra. So that's what Mike's had to install on Tuesday, and it formatted the MBP's 500GB SSD as a single APFS partition—which High Sierra and above inescapably do with any SSD.
    • Tuesday morning I  connected an external Firewire HDD to my 2010 Mac Pro backup server running Retrospect Mac 15.6.1.105 under macOS 10.12.6 Sierra, changed the permissions on all existing top-level folders to Read-Write for all users, and ran the same restore to the HDD.  It ran, but got over 90 cases of "[*] MapError: unknown Mac error 22" and ended with "!Trouble writing folder "/Volumes/Macintosh HD OS X+/.HFS+ Private Directory Data", error -1,101 (file/directory not found) 3/31/19 2:11:03 AM: Execution incomplete".  The HDD remains formatted with HFS+.
    • Tuesday afternoon I took my MBP home, booted it, and—using the pair of adapters I had bought to go USB-C-to-Firewire—ran Migration Assistant to copy my backed-up files from the external HDD onto the MBP's SSD. I chose the iCloud option, because I want to be able to access those files from my MacPro in case my MBP gets messed up again. It wouldn't install SUIDGuardNG.kext, but that turns out to be incompatible with High Sierra.
    • My MBP then would boot under High Sierra, but only using the MBP's built-in display. I normally use an inherited Apple 27-inch LED Cinema Display, connected to my MBP via a KVM switch and a StarTech USB32DPPro adapter. That adapter requires software from DisplayLink.com; at first it wouldn't install. I eventually found this and this Web page, and got the Cinema Display working Tuesday night.

    So what do you suggest as a substitute for this complicated process?  I should note that, having also bought from Mike's a pair of adapters to go USB-C to-Firewire, I have since tried to boot my MBP from my external Firewire HDD.—it doesn't work.

     

     


  21. Here's DovidBenAvraham's 13 April 2019 comment about the Web-based Management Console with its missing-from-the-Configurator Add-On that allows Shared Scripts, as excerpted—with a few square-bracketed clarifications—from the Talk page of the Wikipedia article:

    Quote

    More than two weeks ago [now actually more than 4 weeks ago] two senior Retrospect Inc. salespeople told my friend [me] that the Add-On would soon be included in the Configurator—but that hasn't happened yet. The likely reason is that the Knowledge Base article for Shared Scripts says [here] "Note that as of March 5, 2019, deployment options are limited to ProactiveAI scripts with standard source containers [("All sources", "All local", "All clients", "All network", "All email")] to cloud destinations with simple scheduling options. Support for local sources, local destinations including disk, scheduled scripts, and more extensive scheduling options will be available soon." So the basic capability for deploying shared scripts is officially released, but in Retrospect 16.0 it's so minimal as to be in practice useless for ordinary backup administrators. OTOH aggregation-drilldown within organizations, which also requires the same Add-On, is immediately useful to Partners—the Retrospect Inc. term for consultants who market the backup software to organizations. So it looks to us as if Retrospect Inc. is currently only marketing the license code Add-On to Partners, although a non-Partner administrator can buy it by phoning Retrospect Sales and saying [ ;) ] "pretty please with cherries on top". I [DBA] had considered removing Shared Scripts from the article, but enhanced deployment options are likely to be made available with the release of Retrospect 16.1—which past history shows is likely to happen around 15 May 2019. The only factor causing a longer delay would be if the developers cannot resolve the question of whether the Management Console GUI for defining Shared Scripts should look like the equivalent GUI in the Mac variant or the Windows variant (what they released on 5 March seems to us [DBA and me] more like the Mac variant).

     


  22. Adam Ainsworth and Nigel Smith,

    I've never used block-level incremental backups, which are discussed on pages 206-210 of the Retrospect Mac 14 User's Guide.  "Options" on page 208 says "With block level incremental backup enabled, files 100 MB or larger will be backed up incrementally by default. Smaller files will automatically be backed up in full because restore overhead outweighs the benefits of incremental backup. "  My only database (I'm a home user with a tiny business) is less than 6MB.

    R.V., which I have been forbidden by the head of Retrospect Tech Support from discussing on these Forums, only runs on Windows machines—check the System Requirements.  That worthy (or one of his subordinates) also told me R.V. doesn't even have the concept of a Client.  Think of it as a cheaper competitor to Veeam.  Then instead buy 4 more blank tapes, Adam, as Nigel suggested—IMHO it'd be cheaper and less work to set up .


  23. For reasons, described in this Ars Technica Mac Forum thread, last week I was forced to convert my MacBook Pro's  SSD to APFS.  I therefore, for reasons stated above in this thread, had to upgrade to Retrospect Mac 16.0.1.105.

    Obviously I Removed and re-Added my MBP with Use Instant Scan un-checked.  I now find my SSD scans taking 4 minutes, whereas using Retrospect Mac 15.6.1.105—with the SSD formatted with HFS+ and using Instant Scan—they took 7 minutes.  I also find the total time for an incremental backup of 5GB, including scan, is 14 minutes; using Retrospect Mac 15.6.1.105—with the SSD formatted with HFS+ and using Instant Scan—it took 17 minutes.

     


  24. 1 hour ago, twickland said:

    The peculiar reported capacity of these tapes may be related to the media set locked-unlocked bug I reported here back in 2015 that still has not been fixed.

    twickland,

    You should file a Support Case for that bug, in case the head of Retrospect Tech Support forgot to feed it into their evidently-sketchy bug list.  You'd basically just need to copy the contents of your 2015 Retrospect Bug Reports post into the Case; I'm not posting this suggestion in that thread because I don't know if anybody reads that sub-Forum anymore.


  25. Adam Ainsworth,

    DLT is not "all the same" as LTO.  As the only sentence in the second paragraph of the Wikipedia article says, "In 2007 Quantum stopped developing DLT drives, shifting its strategy to LTO ."

    Make sure that you are really running the Retrospect 14 Engine, not just the Retrospect 14.6.2 Console—which can be run with Engines as far back as Retrospect 12.5.  Page 240 of the Retrospect Mac 14 User's Guide says Retrospect 14.0  "Fixed issue where auto-cleaning request for tape devices was ignored (#6171)".  Pages 49-51 of the UG covers "Cleaning Your Tape Drive".

    Also see this post regarding automating use of a cleaning tape on your tape drive.   For the Retrospect Mac 14 User's Guide, the relevant page numbers are 49-51and 44.

    P.S.: Added 2nd paragraph.

    P.P.S.: Added 3rd paragraph.

×