Jump to content

Don Lee

  • Content count

  • Joined

  • Last visited

  • Days Won


Everything posted by Don Lee

  1. I have noticed that the "disk" media sets have the full path of the members embedded somewhere, and if the data is moved, it is no longer accessible to retro 9.0.1. I did some experiments this morning with a small disk media set, backing up some data to it, then moving it, removing it from the media set list, and doing a "locate". Retro can find the catalog, and seems happy, until I try to restore from the media set. When I do, Retro wants its "member". I have tried all the combinations I can think of to sweet talk Retro into finding the slightly moved members, and I have been unable to get it to work. It appears that once any portion of the path to the data member has changed, Retro is unable to access it. Either some way should be provided to "find" these members, or Retro should warn you not to move them.
  2. Don Lee

    Time zone handling wrong in email

    I checked one more thing: the "date" output on the Mac OS X 10.6 host system where the engine is running: host-213:~ vmadmin$ date -u Wed May 23 16:59:17 UTC 2012 host-213:~ vmadmin$ date Wed May 23 11:59:21 CDT 2012 host-213:~ vmadmin$ This confirms to me that the OS thinks it is "UTC -5". The "Date:" header is part of the body of the message, and is not correct. How the difference in mail servers made testing come out differently is a mystery to me, but the Date header is still not right.
  3. Don Lee

    Time zone handling wrong in email

    I checked a couple more machines. The server (mailhost.icompute.com) is set up correctly, and the timestamps being inserted by the server are showing the correct timezone (UTC -0500). It's the "Date:" header, which is part of the body of the message generated by the Retrospect engine that is incorrect. Example of a message from NetBSD: (note server timestamp agrees with "Date:" header) Return-Path: <nospam@mercy.icompute.com> Received: from mercy.icompute.com ( by mailhost.icompute.com with ESMTP (EIMS X 3.3.9) for <nospam@icompute.com>; Wed, 23 May 2012 11:40:20 -0500 Received: by mercy.icompute.com (Postfix, from userid 1001) id C4A7977013; Wed, 23 May 2012 11:40:18 -0500 (CDT) To: nospam@icompute.com Subject: blah Message-Id: <20120523164018.C4A7977013@mercy.icompute.com> Date: Wed, 23 May 2012 11:40:18 -0500 (CDT) From: nospam@mercy.icompute.com (Donald Lee) time 11:40 AM Now an example from Retrospect: Return-Path: <retro@c.icompute.com> Received: from host-213.icompute.com ( by mailhost.icompute.com with ESMTP (EIMS X 3.3.9) for <retro@c.icompute.com>; Wed, 23 May 2012 11:41:23 -0500 Date: Tue, 23 May 12 11:41:21 -0600 To: retro@c.icompute.com From: retro@c.icompute.com Content-Type: text/plain; charset="UTF-8" Mime-Version: 1.0 Subject: =?utf-8?B?UmV0cm9zcGVjdCBub3RpZmljYXRpb24gZnJvbSB2aXJ0dWUgKDUvMjMvMjAxMik=?= Message-ID: <3420636083-568217244@mailhost.icompute.com> <x-charset UTF-8>This email has been generated by Retrospect to test your settings. </x-charset> Now an example from a linux system. Return-Path: <nospam@vm1.icompute.com> Received: from psmtp.com ( by mailhost.icompute.com with SMTP (EIMS X 3.3.9) for <nospam@icompute.com>; Wed, 23 May 2012 11:49:50 -0500 Received: from vm1.icompute.com ([]) by exprod6mx264.postini.com ([]) with SMTP; Wed, 23 May 2012 12:49:50 EDT Received: by vm1.icompute.com (Postfix, from userid 1000) id C29F362A36; Wed, 23 May 2012 11:49:48 -0500 (CDT) Date: Wed, 23 May 2012 11:49:48 -0500 To: nospam@icompute.com Subject: another test User-Agent: Heirloom mailx 12.2 01/07/07 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Message-Id: <20120523164948.C29F362A36@vm1.icompute.com> From: nospam@vm1.icompute.com (Donald Lee) X-pstn-neptune: 0/0/0.00/0 X-pstn-levels: (S:43.52964/99.90000 CV:99.9000 FC:95.5390 LC:95.5390 R:95.9108 P:95.9108 M:97.0282 C:98.6951 ) X-pstn-dkim: 0 skipped:not-enabled X-pstn-settings: 3 (1.0000:1.0000) s cv gt3 gt2 gt1 r p m c X-pstn-addresses: from <nospam@vm1.icompute.com> [361/14] 11:49 AM CDT Note the Date: header agrees with the linux/postfix-generated timestamp, which agrees with the mail server (mailhost.icompute.com). The Postini mail server says 12:49:50 EDT, but unfortunately, does not have the "UTC -0400" notation. I think the engine doesn't do it right. Sorry to rain on your parade. ;->
  4. Don Lee

    Time zone handling wrong in email

    Hi Robin. I'm confused. This is a "raw" email message from the engine. Engine is running Mac OS X 10.6 on an intel Mini. I'm in Minnesota, and daylight savings time. UTC -5 is correct. Return-Path: <retro@c.icompute.com> Received: from host-213.icompute.com ( by mailhost.icompute.com with ESMTP (EIMS X 3.3.9) for <retro@c.icompute.com>; Wed, 23 May 2012 08:28:40 -0500 Date: Tue, 23 May 12 8:28:40 -0600 To: retro@c.icompute.com From: retro@c.icompute.com Content-Type: text/plain; charset="UTF-8" Mime-Version: 1.0 Subject: =?utf-8?B?UmV0cm9zcGVjdCBub3RpZmljYXRpb24gZnJvbSB2aXJ0dWUgKDUvMjMvMjAxMik=?= Message-ID: <3420624520-568217218@mailhost.icompute.com> <x-charset UTF-8>Script: ToDisk Date: 5/23/2012 Script ToDisk: waiting for media </x-charset> Note that the timestamp inserted by the server is correct - 8:28 AM -0500. The "Date: Tue, 23 May 12 8:28:40 -0600" header in the body of the message looks wrong. I've checked the machine running the engine, and it is also set correctly, as far as I can tell. My mailer is the one flagging this. My mailer is very old, and possibly confused, but the "Date:" header looks like the culprit.
  5. Don Lee

    cannot even buy an upgrade

    For the record, if anyone feels the need to give someone money, I will meet your needs. ;->
  6. I am rebuilding the catalogs on some Retro 6.1 tapes to see if I can restore from them. When the first rebuild was done, I looked at the log: + Executing Rebuild at 4/22/12 (Activity Thread 1) To Backup Set MercyShare10 [001]... !Backup Set format inconsistency (4 at 317871024) !Backup Set format inconsistency (4 at 317768288) !Backup Set format inconsistency (4 at 317960128) !Backup Set format inconsistency (4 at 36464464) !Backup Set format inconsistency (4 at 317530656) !Backup Set format inconsistency (4 at 295383072) !Backup Set format inconsistency (4 at 296098384) !Backup Set format inconsistency (4 at 293092656) !Backup Set format inconsistency (4 at 295433920) ... ... a whole bunch of these..... ... !Backup Set format inconsistency (4 at 314986544) !Backup Set format inconsistency (4 at 315011504) !Backup Set format inconsistency (4 at 315020688) !Backup Set format inconsistency (4 at 315034864) !Backup Set format inconsistency (4 at 315057760) !Backup Set format inconsistency (4 at 315054256) !Backup Set format inconsistency (4 at 315097008) !Backup Set format inconsistency (4 at 315105376) 4/22/12 10:09:19 PM: Execution completed successfully Completed: 3173 files, 39.7 GB Performance: 150.3 MB/minute Duration: 07:16:45 (02:46:45 idle/loading/preparing) What is this "inconsistency" it is complaining about, and then deciding that the rebuild went just fine!?? There were three tapes in this set.
  7. I recently enabled the tape cleaning interval after reading on HP's website that 24 hours of "pulling tape" is the recommended interval between cleaning for my drive (HP storage works DAT72 drive). I find that with this set to non-zero, Retro 9.0.2 will send me email when the tape needs to be cleaned. Cool. Question: Aside from the email, I see no indication that I need to clean the tape drive. If I'm not set up to send email from the engine, how does Retro tell me to clean the drive? Am I missing something on the console? Is there a counter somewhere that tells me how many hours have been racked up?
  8. For the record, I still get lots of these errors ("!Backup Set format inconsistency (4 at 123123123)") when recataloging tapes. It would be nice to know what the error means. It can't be a _good_ thing...
  9. This may well be related to another problem I've been having that seems to be triggered by insertion of a cleaning tape. See it here.
  10. It used to come in the physical package when you bought Retrospect. Take a look at the product page and/or service and see if an "Emergency" CD is still included. If you bought Retro recently, you may be able to get Retrospect to send you one.
  11. The resultant Retro 9 media set (FILE) remains troubled. I copied off the file that was the result of the retro 6.1 recatalog, so I can reproduce this if required. I then did a rebuild on the v9 media set (FILE). The rebuild yielded this: + Executing Rebuild at 5/16/12 (Activity Thread 1) To Backup Set iC_User12 [004]... XcaSixOne::ArxExec (2) skipping empty session tree; rootref is zero XcaSixOne::ArxExec (1) skipping empty session tree; rootref is zero 5/16/12 12:11:00 PM: Execution completed successfully Completed: 184885 files, 47.5 GB Performance: 3,056 MB/minute Duration: 00:15:56 (00:00:01 idle/loading/preparing) I then did a verify on the rebuilt media set. When I do the verify operation, the log looks like this: + Executing Verify at 5/16/12 (Activity Thread 1) To Backup Set iC_User12 [004]... XcaSixOne::ArxExec (1) skipping empty session tree; rootref is zero 5/16/12 12:42:09 PM: Execution completed successfully Completed: 184885 files, 47.4 GB Performance: 4,432.1 MB/minute Duration: 00:10:56 I have saved off the 50 GB file in case anyone wants to look at it.
  12. Try this at home. Retrospect 9.0.2. Mac OS X 10.6.8 (engine host) HP storage works DDS USB drive (DAT72) Write a tape media set that has more than one tape. When you get to the end of the first tape, put a cleaning tape into the drive. When the cleaning is done, inset the next tape. Go to the "activity" window, and click on the little "choose media" button. Select the tape icon under the drive. (drive says "empty") Click OK Wait at least 3 minutes, and watch the SPOD. Suppress the urge to be impatient. Have lunch. Eventually, a dialog comes up saying that the "drive is empty" (it's not) After that, things work fine, the drive sees the tape, etc.
  13. Interesting new behavior. During the "sleep time" between the drive cleaning, and the time the drive/engine wake up again, I saw that the media selection window is blank - showing neither the CD drive, nor tape drive. A screen shot is here.
  14. I have retro 9 client installed on Intel Mac OS X 10.5.8 machine, and the history panel stays blank, even tho many backups have occurred. Progress bar, and other stuff in client looks fine (except for bugs here and here)
  15. I used the Applescript version of the notification script for years, but I had to make some significant changes to get it working properly. Some of the bugs I remember fixing included things like it losing information if ou stopped the backup server, and throwing away information if the output email grew bigger than 32K characters. As I recall, there were a bunch of other bugs, too. (I can post my version if you like - it works with Eudora, and is somewhat special to my setup) I recommend taking a long hard look at the script. It probably got less debugging attention than the Applescript one, and has the same kinds of bugs.
  16. Don Lee

    Retrospect 8.2 is dangerous

    Retrospect 9 is much improved. I recommend it.
  17. Don Lee

    Upgrading Retrospect HELP!

    Yes. This is required. As with other transitions from old formats, the old tapes/sets are read-only. This transition (6.x --> 8/9)also requires re-cataloging the set(s). In theory, simply knowing that you can read the old sets is sufficient. I write a lot of tapes that are *never* read. It is the rare case when I have to dig one up and use it. Vital, but fairly rare. To me, the badness here is that you effectively have to read the backup tape twice to do a restore. It's a pain, but not fatal, IMO. I'm actually going through my tapes and re-cataloging, and converting to retro 9 because retro 9 allows me to "distill" my old backups into single mega-backup sets that take a fraction of the old space by virtue of not re-duplicating the duplicate files when I copy multiple monthly sets into one yearly set. I love being able to do more than one thing at a time w/ retro 9. Yes, Retro 9 has warts. There are bugs to fix, but all in all, it is well worth the time, money, and effort.
  18. Foolishly thinking that updating the client (PowerPC running Mac OS X 10.5.8) from 6.2 to 6.3.029 could be accomplished by updating with the latest updater (9.0.x), I got some strange behavior, until I figured out that when I requested the client update, and selected the client 9.0 rcu file, the result on the client was to simply "turn off" the client. Updating with the 6.3.029 client updater worked fine. The 9.0 client update should fail less disruptively.
  19. In experimenting with the new (v9) client, I figured out that none of the options allowing the client to do anything were enabled, so I went back to the server, and clicked the checkboxes on the appropriate client, then went back to find that all the controls in the client were still grey. I could not turn the client off, I could not add private folders. I could not delay the next backup. I tried rebooting the client machine. No difference. I tried killing and restarting the client with "sudo kill pid pid" and still nothing changed in the client. I don't remember exactly, but I think it was the second time I did the kill, and included both the "retroclient" and quit/restarted the control panels that the controls finally came available to me. It is suboptimal to have those checkboxes on the server that apply to the client with no indication of when the changes will take effect, and fairly clearly a delay. I am left with a question of when the clicking of the checkbox makes a persistent change to the client config. Is it immediate? I would like to see all these sorts of changes have some sort of "OK" or "Apply" button, coupled with some indication of feedback when the changes will take effect.
  20. The new (v9) client allows the server to say whether the client is allowed to turn the client off, or set "private" folders. This is done in the client by setting the controls "grey", and not allowing them to function. If a client is not allowed to select private folders, he can still drag a folder to the "private folders" window, and the folder will be added to the list. The user is then unable to *remove* the folder.
  21. If there is a requirement or preference for the catalog and the disk "member" to be on different volumes, that should be documented. Unfortunately, I ran an experiment that says the "same volume" theory is wrong. I created a disk backup set "atrash", with the catalog on one disk, and the data member(s) on another. I then backed up something small to it. Backup worked fine. "recycle" button available. I copied the catalog to the same disk as the data member(s). I then did a "locate" and "verify" to ensure that Retro had accepted the change. The "verify" failed. (?!) + Executing Verify at 5/11/12 (Activity Thread 2) To Backup Set atrash... !Can't save Backup Set atrash, error -1103 ( write protected) 5/11/12 7:36:36 PM: 1 execution errors Completed: 263 files, 39.8 MB Performance: 2,388.7 MB/minute Duration: 00:00:02 I checked on the catalog file, and it was read-only. I changed it to read/write, and did the "verify" again. Works now. (side question: why does retro insist on having write access to the catalog for a *verify* operation??? It looks like it actually did finish, but the red X says it didn't.) When done, checked the "recycle" button. Available. I pressed it. No more files in "atrash".
  22. OK, I give up. When I select a media set, the "recycle" button is made available to me in the "media sets" panel. I have 3 "disk" sets, and when I select them, the "recycle" is greyed out. Why? I have checked permissions, and that does not seem to be the problem. There is nothing unusual about these sets, except that both the catalog and the member data is on Firewire disk. The single "disk" set I had where the catalog file was on the "main" hard disk (in /Library/Application support/Retrospect) THAT one, I could recycle. Why is this button unavailable?
  23. Interesting notes on tape drives here. Notes of time between cleanings, caution about "non-native" tapes, and duty cycles of drives and tapes. http://h20000.www2.hp.com/bizsupport/TechSupport/Document.jsp?objectID=lpg51254
  24. I note that in the tapes that I have written so far with Retro 9, the capacity of the tape, as reported by Retrospect, is improved by around 30% versus that reported by Retrospect 6.1. This could be how the data size is reported, but Retro 6.1 seemed to report the uncompressed data size, and I expect that Retro 9 does the same. (otherwise my numbers make no sense!) I have a USB HP DDS-5 (DAT72) tape drive, and with Retro 6.1, it was rare for me to get more than a hair better than the "native" data capacity. DDS-4 tapes claim 40 GB, but actually hold about 20 GB uncompressed, and get 40 GB only when compression is maximal. With DAT72 tapes, I would normally get 35 GB. Never more than 37. With Retro 9, I'm seeing 45, or even over 50 GB on a single DAT72 tape. Whatever was done, it's good. ;-> (Was the compression broken in 6.1?) Thanks to Retrospect team.
  25. Yes, they are on the same volume. Is that bad? When going through the documentation, I also note that in theory, I should be able to "repair" moved disk media sets. I'll have to try it....