Eric Hansen
Forum Replies Created
-
Haha, I’m the cheap ass client. Bob, I’m so offended.
i think 450MB/s is an acceptable performance hit, vs. the price of the Cubix chassis. at least for me. i can always grow later.
thanks
e
Eric Hansen – https://www.erichansen.tv
-
in my server, i have 2 ATTO R380 cards in slots 1 and 2, and 2 4port Small Tree ethernet cards in slots 3 and 4. i use a Newertech USB to DVI adapter to connect to a 24″ Dell monitor. no real issues. it can’t do anything graphics intensive, but it’s a server, so it works fine.
if you ever need to do a OS reinstall, which i just did last week, you have to put the original GFX card back in. after it’s setup again, you can pull the GFX card.
e
Eric Hansen – https://www.erichansen.tv
-
i’ve been running this config on a server for well over a year. no issues. the only suggestions that ATTO gave me were to make sure that the cards both had the same flash.
we copy files back at forth at full speed with no issues.
e
Eric Hansen – https://www.erichansen.tv
-
hey Mark
it sounds like you’re asking 2 different questions and Bob and Andrew addressed the archive question. i agree with them 100%. you should transfer your camera raw, whatever that is, to something like LTO right away. always save your camera’s raw format, no matter if you transcode for your editing workflow or not.
but to answer your question about codec. you have a few different options. i find it easiest to get everything into a common codec. this is because of FCP7’s need to render anything that doesn’t perfectly match the timeline codec. to keep my editors working and not rendering, i transcode everything to either ProResHQ or ProRes Proxy, all 1920×1080 23.98fps.
we use PR Proxy as an offline codec for very large projects. we just edited a movie using Proxy because it would have required about 200TB of SAN space to be edited at PR HQ, and we only have 56TB. then we swapped to the PR HQ files during the online.
your other option is to deal with the rendering from different codecs, or use plug-ins like those from Calibrated Q.
the ideal is getting to the point where the edit system is codec-agnostic and will do whatever hardware rendering is required on-the-fly. Adobe is close with their Mercury Engine.
i understand there are cases where your footage will live on your SAN forever. for things like stock footage or projects that you constantly revisit. i typically keep the final exports of every project on our SAN, so they can be watched quickly when new clients come through. but SANs are not designed for archive and it’s very expensive for that use. use LTO tapes or something similar for your archive. make sure everything that’s on your SAN, is also backed up somewhere else.
e
Eric Hansen – https://www.erichansen.tv
-
hey Graham
did you update your OS from Leopard to Snow Leopard? i know that catalogs have to be updated by Tolis to make this change.
it sounds like BRU might be writing the new archive catalogs to a different directory, or there’s a permissions issue. either way, call Tolis and they can help you out. i find whenever i have issues with PE it’s a lot easier to give them a call than to try and troubleshoot it myself.
e
Eric Hansen – https://www.erichansen.tv
-
[Bob Zelin] “But I cannot stress enough to you that a direct attached LTO is PROCESSOR INTENSIVE, and if you are doing an LTO backup with the ProMax, or HP Ultrium, or Quantum LTO, while you are doing file sharing, it will BOG DOWN YOUR SYSTEM, and you will get drop frames”
Bob, I have to disagree with you here. for the last year i’ve been running 2 HP LTO-4 drives on a Mac Pro server with 2 ATTO R380 cards (one card is attached to 2 16 drive RAIDs and the other card is attached to another 16 drive RAID and the HP drives). in my experience BRU PE uses up to 50% of one core of the 8 core Mac Pro, and that’s being generous. the LTO-4 drives would pull anywhere from 80-150MB/s, which would barely touch the 600MB/s of headroom the RAIDs had. never got drop frames.
e
Eric Hansen – https://www.erichansen.tv
-
this is how i have my MacBook Pro set up. SSDs will change your life. my 4 year old MBP is faster than the 1 year old Mac Pros in our shop.
e
Eric Hansen – https://www.erichansen.tv
-
make sure that “Put the Hard Disks to Sleep When Possible” is UNCHECKED in System Preferences > Energy Saver
e
Eric Hansen – https://www.erichansen.tv
-
hey Matt and Jeremy
thanks for the suggestions. we ended up going back through the Autosaves and pulled a version that didn’t have the “Unspecified Path” problem. since the project has been added to since, it looks like the AE has about a day of work to update the project, so not as bad as i was fearing.
ok, so this is only a guess as to why this happened: i think media went offline while the projects were open. it’s my understanding that FCP is always checking to see if media is online or moved (there’s a check box for this). which is why FCP will sometimes beach ball if you return to it after being in Finder or another app for awhile. it’s double-checking that all the media is still available. if media is moved, FCP will quickly reconnect to the new media if it’s easy to find. older versions of FCP just didn’t do this. but i think if the media just goes offline and there’s no new media to connect to, or it’s not easily found, FCP will replace the “Source” metadata with nothing. you can see the “Source” metadata by right clicking on a clip and selecting Item Properties. the problem is that there’s no way to replace the information in the “Source” metadata once it’s gone.
FCPreconnect looks like a great app, but didn’t work in this particular situation because the problem wasn’t the media, it was this “Source” metadata in FCP. roundtripping XMLs would add the “Source” metadata back in, but it would be based on FCP’s clip names and not Finder names, which would mess up Subclips. (in our naming scheme, the subclip for Master clip – FLT-110401-Jackson-CF01-01 would change to FLT-110401-Jackson-CF01-01 Subclip 10). so reconnecting is impossible because Subclip names don’t appear in Finder.
so moral of the story (i guess) is don’t delete media while FCP is open.
i’m bummed i couldn’t find a real solution though. i foresee this being a huge problem as we start team editing this movie. i think roundtripping XMLs could work if there was a bulk way to change Subclip names back to their Master clip names in the XML (reading the XML, i couldn’t believe how LONG each clip’s entry was – a bin of 10 shots churned out about 20 pages of code). i should say that roundtripping XMLs does work if you don’t have Subclips. but a lot of our media is either full HDCAM tapes or hour-long interviews on P2, so we have to deal with subclips quite a bit.
thanks
e
Eric Hansen – https://www.erichansen.tv
-
[Ben Watson] “Please don’t stone – I’ve been using Blu-ray burner. Cost has come down, and with 50 gb (double sided) of storage, it’s a pretty cost effective solution. I’m not archiving for decades, just a year or two at most, so it works pretty well for us. Most of my projects will fit on a couple discs or even one depending on the amount of shooting…”
no stone throwing here. archiving to Blu-Ray can work great. here are the 4 main reasons why i don’t use it:
1. not sure how long the discs will last. i have “archive quality” CD-Rs that are no longer readable.
2. i’m often working with files that are bigger than the discs. how does the app you use handle “spanning”?
3. the drives are achingly slow
4. i would need quite a few spindles to archive just my own personal projects. so their footprint is bigger than LTO tapes of equivalent TBsbut if these things are not a concern for you, and they probably aren’t, then archiving to BD is awesome. i have been thinking about moving my personal projects from LTO to BD, just so i don’t need to use my office LTO drives. but these are a backup to bare hard drives on my shelf, so i only need to access them if my bare hard drives go down. LTO tapes take up less space in my safe deposit box than BDs would.
e
Eric Hansen – https://www.erichansen.tv