Jump to content

Doug_M

Members
  • Content count

    16
  • Joined

  • Last visited

  • Days Won

    1

Doug_M last won the day on May 19

Doug_M had the most liked content!

Community Reputation

2 Neutral

About Doug_M

  • Rank
    Member

Recent Profile Visitors

The recent visitors block is disabled and is not being shown to other users.

  1. SOLUTION - rebuilding the catalog file has fixed the problem. Note: Fast catalog rebuild option was enabled during archiving, so I was able to rebuild from the last tape in my media set. Be aware that Retrospect will ask for another tape (last number plus 1, which doesn't exist (!!)). This is Retrospect's way of telling you the rebuild is complete and you should hit the Stop button. Restore of the volume is underway!
  2. I've not had any issues with Disk Utility, aside from the reduced functionality vs. older versions... aside from the disappointing feature reduction, it's been my experience that the disk formatting functions work perfectly well. Having formatted hundreds of drives over the past few years, I feel fairly confident that my tools and methods are appropriate and effective for the task at hand. I singled out GUID because that's the partition scheme for the disk, followed by the volume / filesystem format. In Disk Utility, you have 3 different partition scheme options - 1) Apple Partition Map (for really old Macs, pre-OSX) - 2) Master Boot Record (for PC formatting / FAT16, 32. ExFAT) - and 3) GUID Partition Map (for all modern Macs) which supports HFS, HFS+ (Journaled or not), and APFS in all its variants. Nothing unusual or out-of-the-ordinary for the partition scheme or file system format for the volume in question. As noted above, to avoid any potential issue with the RAID, I removed it from the equation several days ago and instead have been attempting restore with a single 4TB drive in the same enclosure that was used for the original backup.
  3. Central US, so where I am it's almost time to head for the pub. Cheers!
  4. I'll need to extract the machine to disable SIP (it's headless, in the rack, in a closet), but will give it a go. RetrospectInstantScan and RetrospectEngine both have full disk access, and appear in the Files and Folders tab of Security prefs. So far, I've only tried restoring the entire volume, but will give files & folders a try next. EDIT: Same error. SIP is next, though there aren't any 'protected' files on that volume - it's all video clips and project files, so I'm not optimistic on this front. Starting to think I'll need to repair / rebuild the catalog for that set.
  5. Not much to look at, but here you go: + Restore using Restore 0015 at 5/13/21, 4:40:39 PM (Activity Thread 1) To volume ANOD0015... Warning: volume ANOD0015 has the "Ignore ownership" setting enabled. Note: OS X System Integrity Protection may prevent the restoration of protected files and folders to ANOD0015. - 5/13/21 4:40:39 PM: Restoring from Video Bkup, Snapshot /Volumes/ANOD0015, 9/18/18 6:30:09 PM 5/13/21 4:40:41 PM: Selector "All Files" was used to select 12,748 files out of 12,748. !Trouble matching Video Bkup to ANOD0015, error -1,124 (invalid file system data) 5/13/21 4:40:44 PM: Execution incomplete
  6. Thanks guys... I format every new drive prior to use, and I only use APFS on my personal SSDs, or internal boot SSD on newer Macs that require it. The drive in question is a spinner, formatted with Disk Utility as I said earlier - GUID / HFS+ journaled. APFS is a complete non-starter on spinners as far as I'm concerned. I've swapped the 4TB RAID0 destination disk with the single near-line storage unit intended to replace the damaged drive. It's a recent SATA III, unused (but formatted), and named the same as the drive it's replacing. Same error as before. DiskWarrior doesn't have the capability to format a drive - I've been using it for years to optimize or repair damaged directories, which it excels at.
  7. Attempting to restore an entire disk (4TB / GUID / MacOS Extended (Journaled)). The disk in question is a non-booting media drive which was dropped by a careless individual and no longer functions, but was completely backed up a couple of years ago. Running Retrospect 16 on a trashcan MacPro with Catalina, connected to a Tolis 8 slot library (HP drive / LTO6) via HighPoint SAS card in a Sonnet TB2 expansion chassis. My restore destination is a 4TB RAID 0 in an OWC hardware TB2 enclosure, which is blank and is named the same as the dead drive. This is the same rig that created the original backup. Upon running the script, I get the standard warning about ignoring ownership, then Restoring from <media set>, snapshot <volume> X !Trouble matching <media set> to <volume>, error -1,124 (invalid file system data) Execution incomplete. The script runs for just a few seconds before it incompletes. Could this be a volume type mismatch - single spinner vs. hardware RAID? They're almost identical in size (RAID is a tiny bit larger). TIA --Doug
  8. I gotcha. Okay, so I'm getting "Share_Name' for Name, and the server's IP address under Machine. So it looks like OS-level mount. Side note: I get 'Unknown' for the shares in the OS column. Here's an example folder name that gets mangled: AFP - /Animations (Do we need to rerender smaller?)/ SMB - /AVFA3C~5/ I'm not entirely certain that the mangling is the reason for the -1101 errors though. EDIT: In the logs, the only files that were tagged as missing were in some folder with a mangled name (we have many). FWIW.
  9. Hey Nigel, Fortunately, it was after midnight when the tornado went through and was only property damage. We fared much better than most of our neighbors, though. I did try mounting the share first, but it still showed as such in the Sources pane. I've been using the IP address for these connections, which doesn't seem to improve my situation. I tried using SMB again, and that seemed to get further into the process, but there are still major issues with name mangling and files not being found after the initial scan. I get the impression that SMB is functioning more robustly than AFP, but until I can resolve the mangling issue it's a non-starter.
  10. Resurrecting this old thread - our office was hit by a tornado just after midnight March 2. Many temporary workarounds have been in use and we have moved. Just getting the hardware back online - new switch, new Synology box, same data and shares - and (perhaps unsurprisingly) the problem persists. I've been running a test set for troubleshooting. Retrospect is latest v16 (haven't looked at upgrading to 17 yet), trashcan is on Catalina 10.15.6, Synology DSM is current, data scrubbing will occur quarterly and was last run after the transfer to the new NAS. This really feels like a network stack bug or similar, but only happens with Retrospect. Back to the logs!
  11. Hey David! It's an old installation, and had been working reliably. Things got weird when the Retrospect Server machine was updated to Catalina / Retrospect Mac 16 (neither of those updates were my decision, in case you're wondering ). All of this happened over the New Year break, and I stepped in to sort it out last week. I'm fully aware that there aren't any employees on these boards, but there are some knowledgeable folks lurking about... and since the 30/45 day support thing has expired, I thought I'd try my luck. Thanks for the links!
  12. I'll give the local volume backup a try. I've not been using SMB due to name mangling, which is partly due to an smb.conf setting on the NAS, and years of Mac users not following our file/folder naming protocols. That, coupled with the default performance hit due to signing, is why I'm pursuing AFP. Additionally, the volumes in question are on our live production server, so testing some of these changes would require more down-time than I can get. Here's what I'm seeing in Console.app during the hang / reconnect: AFP_VFS afpfs_DoReconnect: Restoring session /Volumes/Test_Volume ASP_TCP ReplayPendingReqs: replaying slot 122 with reqID 5967 afpCmd 72 on so 0x62f0e214331bf2ef ASP_TCP ReplayPendingReqs: replaying slot 123 with reqID 5968 afpCmd 72 on so 0x62f0e214331bf2ef ASP_TCP ReplayPendingReqs: replaying slot 124 with reqID 5969 afpCmd 72 on so 0x62f0e214331bf2ef AFP_VFS afpfs_DoReconnect: get the reconnect token ASP_TCP CancelOneRequest: cancelling slot 83 error 32 reqID 29780 flags 0x9 afpCmd 0x40 so 0x62f0e214331bf2ef ASP_TCP Disconnect: triggering reconnect by bumping reconnTrigger from curr value 3993 on so 0x62f0e214331bf2ef AFP_VFS afpfs_DoReconnect: GetReconnectToken failed 32 /Volumes/Test_Volume AFP_VFS afpfs_DoReconnect: doing reconnect on /Volumes/Test_Volume AFP_VFS afpfs_DoReconnect: soft mounted and hidden volume so do not notify KEA for /Volumes/Test_Volume AFP_VFS afpfs_DoReconnect: Max reconnect time: 30 secs, Connect timeout: 15 secs for /Volumes/Test_Volume AFP_VFS afpfs_DoReconnect: connect to the server /Volumes/Test_Volume Ignoring source 0x7fdf9641ca40, snapshot is from kernel kernel_task TCP4 flow id 44472 (close) so pkts rx 6 tx 9, bytes 456 845 cell 0 0 wifi 0 0 wired 696 1205 deltas 456 845 0 0 0 0 696 1205 AFP_VFS afpfs_DoReconnect: Logging in with uam 8 /Volumes/Test_Volume AFP_VFS afpfs_DoReconnect: Restoring session /Volumes/Test_Volume These are generated by 'kernel', and repeats several times every second - it's a flood of reconnect attempts. Funny thing is that everyone here uses AFP, most are running Catalina, and we have no issues with availability except in this instance. And here's what I'm seeing in my Retrospect test script log when it actually makes it through (rare): + Normal backup using NAS TEST at 2/21/20, 4:13:48 PM (Activity Thread 1) To Backup Set TestSet... - 2/21/20 4:13:48 PM: Copying Marketing_Sales [*] 2/21/20, 4:17:53 PM: Not done after 30s: [*] 0:/Volumes/Marketing_Sales/Production for Proposals/2018/JOB 18-1997/01 Accounting/ [*] 1:/Volumes/Marketing_Sales/Production for Proposals/2018/JOB 18-1997/Proposal/ [*] 2:/Volumes/Marketing_Sales/Production for Proposals/2018/JOB 18-1997/00 Project Management/ [*] > 3:/Volumes/Marketing_Sales/Production for Proposals/2018/JOB 18-2101/00 Project Management/ [*] 2/21/20, 4:20:55 PM: mvlcProcessScanNodes resumed after 212s *File "/Volumes/Marketing_Sales/Production for Proposals/Other Folder”: can't read, error -1,101 (file/directory not found) *File "/Volumes/Marketing_Sales/Production for Proposals/Yet Another Folder“: can't read, error -1,101 (file/directory not found) 2/21/20 4:20:58 PM: Found: 19188 files, 8200 folders, 198.2 GB 2/21/20 4:20:59 PM: Finished matching *File "/Volumes/Marketing_Sales/Production for Proposals/2018/JOB 18-1997/01 Accounting/Invoices/": can't read, error -1101 (file/directory not found) and 8,297 others 2/21/20 4:20:59 PM: Selector "No Files" was used to select 0 files out of 19,188. 2/21/20 4:20:59 PM: Copying: 0 files (0 bytes) and 0 hard links 2/21/20 4:21:02 PM: Building Snapshot... !Can't read state information, error -1,101 (file/directory not found) 2/21/20 4:21:02 PM: Execution incomplete Duration: 00:07:14 (00:07:12 idle/loading/preparing)
  13. Hey David! I saw that, too. I've had Instant Scan disabled in System Prefs/Retrospect pane, though it does have Full Disk Access permission. As the NAS can't run Client in a sanctioned way (I'm not installing the Linux client), I was under the impression that Instant Scan would simply ignore that volume regardless. I switched it off anyway, with no apparent difference. P.S. @Nigel Smith - the alert is a simple system alert dialog with yellow triangle (!) "Server is shutting down. The file server "mySynologyNAS" is shutting down. [OK]
  14. One other detail - the only thing between the two units is a network switch. They live next to each other in the rack. The Synology and switch are set up for Dynamic Link Aggregation (IEEE 802.3ad LACP) bonding all four ports on the 2414, and functions perfectly in every other scenario. Looking at the 'Past Backups' tab, the Status alert says Backup Interrupted.
  15. Thanks for the replies! Full Disk Access is indeed enabled for both the Engine and Instant Scan, the Synology doesn't sleep, and neither does the trashcan. These are scheduled scripts, but I have kicked them off manually to test. The only other thing going on with the Retro server is a time machine backup of the boot drive, to another share on the same fileserver, which I've disabled temporarily for testing. Everything in the scripts is default. I've checked the Console, and logs in Catalina are now very different and in this case useless, the logs on the Synology don't show anything unusual... successful connections to the specified shares, but nothing about disconnects. The maddening thing is that I only get the last 500 lines of the Retro log, and I'd like to see the whole thing. What I can see is mostly this: *good path to actual file/filename - can't read, error -1,101 (file/directory not found) And then the script closes with this: 2/20/20 7:38:15 PM: Building Snapshot... !Can't read state information, error -1,101 (file/directory not found) 2/20/20 7:38:17 PM: Execution incomplete No longer present: 4,117 files, 219.2 GB Completed: 10 files, 0 B Performance: 0 MB/minute Duration: 01:38:14 (01:37:41 idle/loading/preparing)
×