Author Topic: alternative to filehamster? (Read 109562 times)

urlwolf · « **on:** September 07, 2009, 12:09 PM »

What's a good alternative to filehamster?
it crashed on my windows server 2008 64 machine, and it was a tad too pushy to get the pro version anyway...

tomos · « **Reply #1 on:** September 07, 2009, 02:53 PM »

did you see this thread ?
AJC Active Backup ---- at Bits du Jour

work your way down from this post
Re: MOGware's FileHamster (Complaint)
but hey I see you posted there

tranglos · « **Reply #2 on:** September 07, 2009, 07:42 PM »

did you see this thread ?
AJC Active Backup ---- at Bits du Jour
-tomos (September 07, 2009, 02:53 PM)

AJC Active Backup is certainly not as memory hungry as FileHamster (and snappier too), but if you read the AJC forum, a new beta version is about to be released, and it's written in .Net now, just like FH. I fear the worst!

The good thing about FH is that it stores each backup file as a physical file, with a timestamp, so restoring is a breeze. AJC uses a proprietary file format, so you need the (free, separate) viewer app to extract files from the compressed backup files. That said, I took the BdJ offer and have switched to AJC, since the memory footprint of FileHamster is quite obscene, and it has a few other quirks I wasn't happy about. I don't expect to be upgrading when AJC switches to .Net though $:-\$

mwb1100 · « **Reply #3 on:** September 07, 2009, 09:55 PM »

JP Software (makers of 4NT/Take Command) have a product call Cascade Point that is a a similar automatic backup that saves file revisions. I haven't used it, so I'mm not sure how well it works, but may be worth a try:

http://www.jpsoft.co.../cascadepointdes.htm

mwb1100 · « **Reply #4 on:** September 07, 2009, 09:58 PM »

but if you read the AJC forum
-tranglos (September 07, 2009, 07:42 PM)

Do you have pointer to the AJC forum? It doesn't appear to have a link from their website.

tranglos · « **Reply #5 on:** September 08, 2009, 04:36 AM »

Do you have pointer to the AJC forum? It doesn't appear to have a link from their website.
-mwb1100 (September 07, 2009, 09:58 PM)

You're right, there is no forum! Sorry about he confusion. I remembered reading that comment on a board - that board was Bits du Jour comment section:
http://www.bitsdujou...ive-backup/#comments

The author's response is Aug 21 2009 5:47am

elpresi · « **Reply #6 on:** September 08, 2009, 10:56 AM »

I have not used it, but AFAIK Docshield is similar and free for non-comercial use:
http://www.docshield.com/

Carol Haynes · « **Reply #7 on:** September 08, 2009, 11:04 AM »

DocShield requires Microsoft .NET framework version 2.0 running under any Windows operating systems capable of running the .NET framework This includes Windows Vista, Windows XP, Windows 2000, Windows 98, Windows ME, and Windows NT.

mwang · « **Reply #8 on:** September 08, 2009, 11:31 AM »

Genie-Soft (the company behind Genie-Soft Backup Manager) has a new product called Genie Timeline.

The most interesting part is its tiered backup scheme. It makes a backup every half hour, then it "saves the hourly backups for the past 24 hours, daily backups for the past month, and weekly backups for everything older than a month." (quoted directly from its user manual.)

It's something I've always wanted, and has been available to Mac users for a long time with the famed Time Machine. In fact, I believe Genie-Soft got its idea from Apple, with the similarity in names and the feature comparison table between the two products on their web site.

I downloaded the trial version yesterday and tried it briefly. It worked as promised, and wasn't taxing the system too much (didn't slow down my computer during the few hours of trial). It's designed to be easy to set up at the expense of flexibility, however, which bothered me a lot, so I uninstalled it after a few hours. They have a page comparing Genie Timeline to Backup Manager Pro, which should give you an idea of its strength and weakness. The user manual (pdf) can also be downloaded.

I now have my own tiered backup scheme, using SFFS, DOpus, Powerpro and batch files. If you don't need a "tiered" system (I believe FH doesn't have one), just need something that sits in the background and back up frequently and quietly, then SFFS, SyncbackPro, MirrorFolder and some others should do it, though each with its own strength and weakness. But I think you know them all, since they've been mentioned many times here.

mwang · « **Reply #9 on:** September 08, 2009, 11:39 AM »

I tried everything mentioned above, but didn't keep any of them. FWIW, here are some terse comments from my notebook:

AJC Active Backup 1.5.10 (decent, with two tiers, but I want more tiers and better looking interface)
CascadePoint 2.10 Build 29 (need to set up file types)
DocShield v.2.0.2 (no wildcard support for folder exclusion)
MirrorFolder 4.1.192 (can't preview what's included and what's not. not flexible in file/folder exclusion)

mwb1100 · « **Reply #10 on:** September 08, 2009, 12:04 PM »

I now have my own tiered backup scheme, using SFFS, DOpus, Powerpro and batch files. If you don't need a "tiered" system (I believe FH doesn't have one)...
-mwang (September 08, 2009, 11:31 AM)

FH does support a 'tiered' set of revisions (not sure if it's in the free version, though). They call it "DynamicRevisionHistory" and it's buried in the options for a 'watch'. It seems to be on by default (but maybe I set mine - all I know is that it's on right now for me), and "keeps all revisions from the previous 24 hours then only keeps the latest revision for each previous day, each week, month and finally year."

mwang · « **Reply #11 on:** September 08, 2009, 12:19 PM »

FH does support a 'tiered' set of revisions (not sure if it's in the free version, though).
-mwb1100 (September 08, 2009, 12:04 PM)

Interesting. I'll have to take another look at FH, then. Thanks.

Carol Haynes · « **Reply #12 on:** September 08, 2009, 12:54 PM »

You could have a look at AutoSave Essentials. I used to use AutoSave 2 but development seemed rather slow (there have been no new updates in at least 2 years to AutoSave 2) and I moved to FileHamster. Not sure what the difference is between AutoSave 2 and Essentials but it seemed to do what it said on the tin (which includes file versioning backups).

tomos · « **Reply #13 on:** September 08, 2009, 02:35 PM »

FH does support a 'tiered' set of revisions (not sure if it's in the free version, though). They call it "DynamicRevisionHistory" and it's buried in the options for a 'watch'. It seems to be on by default
-mwb1100 (September 08, 2009, 12:04 PM)

yes, it's on by default in paid version - even for backups already setup - I cant believe it - it's newish I think,
that's the last thing I want

I've posted in FH forums asking for more info but the help file says:

Dynamic Revision History when set to True will clean your revisions for you based on what day they were made on. This will help to keep your library from being overfilled with a large amount of old revisions.

First day - All revisions will be kept.
Next day and each day for one week - One revision per day for the week
Next week - One revision per week for the same month
Next month - One revision per month up to the end of the first year
Next year - One revision per year

f0dder · « **Reply #14 on:** September 08, 2009, 05:52 PM »

Oh, is Genie Timeline available now? I looked a couple of weeks ago, and at least I couldn't find a download link back then. Would love to play around with it.

It's not like what apple's doing with Time Machine is that new or unique, by the way - it's been possible to do with rsync + scripting for quite a while. Apple just pretty-GUI'ed it

. Afaik Time Machine does a full file backup whenever a file is changed, Timeline has an advantage there since it does block level backups and thus captures only the changed parts...

cranioscopical · « **Reply #15 on:** September 08, 2009, 05:58 PM »

Here's an alternative: 12Ghosts Backup.

"HyperBackup automatically stores multiple versions of your files and applies a reasonable filter to keep versions of every stage."

f0dder · « **Reply #16 on:** September 08, 2009, 07:23 PM »

I had a quick look at Timeline, and it doesn't seem like it's installing any drivers - I'm wondering how they're doing block-level backups in an efficient way if they don't have a filesystem filter driver to record the changes?

mwang · « **Reply #17 on:** September 08, 2009, 07:28 PM »

It's not like what apple's doing with Time Machine is that new or unique, by the way - it's been possible to do with rsync + scripting for quite a while.
-f0dder (September 08, 2009, 05:52 PM)

rsync is great. It's what I use on my linux server, but I don't know how to do it with Windows boxes. DeltaCopy had no unicode support the last time I checked.

Apple just pretty-GUI'ed it . Afaik Time Machine does a full file backup whenever a file is changed, Timeline has an advantage there since it does block level backups and thus captures only the changed parts...
-f0dder (September 08, 2009, 05:52 PM)

Timeline does block level backup only for files larger than 40MB, according to its manual. That's not much use for me (I don't use Outlook). SpiderOak uses that technology to the extreme, on the other hand. (As an online backup provider, it stands to gain much more from it.)

And yet my current local backup solution doesn't do delta copy at all, so it's not a minus against Timeline for me. What really drove me away was the lack of customizability, in what files to back up, etc. It's a new product after all, so hopefully it'll be more flexible as it matures.

mwang · « **Reply #18 on:** September 08, 2009, 07:42 PM »

I had a quick look at Timeline, and it doesn't seem like it's installing any drivers - I'm wondering how they're doing block-level backups in an efficient way if they don't have a filesystem filter driver to record the changes?
-f0dder (September 08, 2009, 07:23 PM)

No, it didn't install any driver (inside the Windows directory tree if that's what you mean) according to my log. It has a "BlockLevel.dll" and it puts a ton into the registry. Do you really need a "driver" to do block-level backups? It registered as a set of system services and does its work in the background. Can't it perform block-level backups that way? That seems to be how SpiderOak does it as well.

f0dder · « **Reply #19 on:** September 08, 2009, 07:49 PM »

mwang: there's rsync version(s?) for Windows, the one I bumped into was a cygwin build... and I try to stay away from cygwin as much as possible. rsync in and by itself isn't enough for backup though, imho - it should be combined with script/whatever to get the same kind of functionality as time machine has. I use rsnapshot on my linux box, works like a charm.

Block-level backups are cool, but I don't really need it for the stuff I want backed up. Also, without a filter driver (or something really smart I don't know about), I dunno how they can detect block changes efficiently (comparing to a backed-up version would be stupid). I was mostly excited about the block-level feature because I thougth it would mean a filter driver (I've got a thing for those

)... which would mean the backup app instantly knowing about file changes in a pretty efficient way.

Haven't had time to play around with Timeline, but it does sound like it's configurability is on the low end - that kinda sucks. And wtf is up with including a web server as part of the application? O_o - takes up ~28MB installed, and seems like a pretty pointless thing to do. Oh, and if explorer is the only interface for restoring backups, then... ugh.

No, it didn't install any driver (inside the Windows directory tree if that's what you mean) according to my log. It has a "BlockLevel.dll" and it puts a ton into the registry. Do you really need a "driver" to do block-level backups? It registered as a set of system services and does its work in the background. Can't it perform block-level backups that way? That seems to be how SpiderOak does it as well.
-mwang

Well, for what I know, you need a filter driver (.sys) to detect "X bytes written too ffset Y in file Z" change notification (rather than just getting "file has changed") - unless you can ask Volume Shadow Copy Service or something. If you don't get the "X,Y,Z" notification, afaik you'd have to do a manual compare of the file to a previous version in order to detect modified blocks. This would work, and conserve space, but it's a lot slower than simply writing out blocks you've been previously told were modified.

I noticed blocklevel.dll as well and took a quick look at it, and it's definitely not a driver. Seems to use SQLite for something, possibly to store list of changed blocks?... but I didn't look any deeper than that.

mwang · « **Reply #20 on:** September 08, 2009, 08:52 PM »

mwang[/b]: there's rsync version(s?) for Windows, the one I bumped into was a cygwin build... and I try to stay away from cygwin as much as possible.
-f0dder (September 08, 2009, 07:49 PM)

That, and DeltaCopy were what I found the last time I looked around. From memory they all had problems dealing with unicode file/folder names, so I gave up.

rsync in and by itself isn't enough for backup though, imho - it should be combined with script/whatever to get the same kind of functionality as time machine has. I use rsnapshot on my linux box, works like a charm.
-f0dder (September 08, 2009, 07:49 PM)

My linux skill is basic, so I followed a tutorial to set up rsync, I believe. Don't really remember what, if any, other tools were employed. I'll keep your advice in mind when I maintain it next time. Thanks!

Well, for what I know, you need a filter driver (.sys) to detect "X bytes written too ffset Y in file Z" change notification (rather than just getting "file has changed") - unless you can ask Volume Shadow Copy Service or something.
-f0dder (September 08, 2009, 07:49 PM)

Thanks for the explanation. I'm not a programmer, so I can only observe. SpiderOak doesn't install a driver either (no .sys files). It does detect block level changes not only because they advertise the feature on their web site, but also the uploads (of changed large files) take so little time that I don't think it's possible otherwise.

It couldn't compare a file to its earlier versions since they sit remotely (well, I do have local backups, but SpiderOak wouldn't know where to look). I don't know if it asks Volume Shadow Copy Service or other system services for the necessary information. But as a cross-platform service with linux and mac os clients, it would have to ask different system functions on different systems if it doesn't do it by itself. That's still possible, of course.

mwang · « **Reply #21 on:** September 08, 2009, 08:57 PM »

Here's an alternative: 12Ghosts Backup.
-cranioscopical (September 08, 2009, 05:58 PM)

Tried this one as well (v. 9.0). No unicode support.

f0dder · « **Reply #22 on:** September 09, 2009, 01:05 AM »

Thanks for the explanation. I'm not a programmer, so I can only observe. SpiderOak doesn't install a driver either (no .sys files). It does detect block level changes not only because they advertise the feature on their web site, but also the uploads (of changed large files) take so little time that I don't think it's possible otherwise.
-mwang (September 08, 2009, 08:52 PM)

Well, when remote uploading is involved, the time saved even on fast ADSL links would make compare-to-get-changes quite acceptable

It couldn't compare a file to its earlier versions since they sit remotely (well, I do have local backups, but SpiderOak wouldn't know where to look). I don't know if it asks Volume Shadow Copy Service or other system services for the necessary information. But as a cross-platform service with linux and mac os clients, it would have to ask different system functions on different systems if it doesn't do it by itself. That's still possible, of course.
-mwang (September 08, 2009, 08:52 PM)

Hm, interesting.

One solution I could think of would be to store checksums/hashes of each block of a file locally, for large enough blocks that this doesn't cause massive data bloat. It would be faster than comparing to an old version (only one big datastream read + the much smaller list of block-hashes). Still requires processing the entire file though.

mwang · « **Reply #23 on:** September 09, 2009, 09:33 PM »

One solution I could think of would be to store checksums/hashes of each block of a file locally,
-f0dder (September 09, 2009, 01:05 AM)

Indeed that could be the case. SpiderOak maintains a huge "pandora_sqliite_database" (117 MB at the moment). In addition there are several .db files, mostly small, but the largest one--snapshot.db--is 52 MB large.

f0dder · « **Reply #24 on:** September 10, 2009, 03:57 AM »

You could grab sqlite.exe / sqlite3.exe from here and have a look-see at what tables those files contain

(note: make a copy before opening, I don't know if sqlite3 will attempt to upgrade an older db on open, or only when modifications are done - but it'd suck if SpiderOak uses v2 format and suddenly the database is in v3 format

).