Chat log from the meeting on 2021-09-28

From OpenSimulator

Jump to: navigation, search
[11:01] Gavin.Hird @grid.xmir.org:8002: Hi Andrew, Jamie
[11:01] Andrew Hellershanks: Hello, everyone.
[11:02] George Equus: Hi Andrew
[11:02] Jamie.Jordan @grid.kitely.com:8002: Hi everybody
[11:02] Ubit Umarov: osg is not using still another proprietary assets service
[11:02] George Equus: Lots of stuff broke this time...
[11:02] Ubit Umarov: and as bad as previus one
[11:02] Selby.Evans @grid.kitely.com:8002: Hi Andrew, Jamie
[11:02] Vincent.Sylvester @hg.zetaworlds.com:8002: With the whole utf8 thing going on I was planning to see if there are ways to use new sql in the connectors to perhaps make them faster, but that would likely discount older mysql or mariadb versions so ultimately would need version detection of the database which doesn't really make sense to implement on such simple functions
[11:03] Ubit Umarov: well you need to report those issues on osgrid foruns
[11:03] Vincent.Sylvester @hg.zetaworlds.com:8002: sql is a lot of fun though now that I am diving head first into it
[11:03] Jamie.Jordan @grid.kitely.com:8002: hi :-)
[11:03] Andrew Hellershanks: Vincent, wait until you get in to using JOIN. :)
[11:04] Ubit Umarov: well seems mariadb is less reliable now than mysql
[11:04] Ubit Umarov: or so i was told
[11:04] Vincent.Sylvester @hg.zetaworlds.com:8002: Way past that already, doing complex string operations and references across databases and tables
[11:04] Vincent.Sylvester @hg.zetaworlds.com:8002: I have been using multiple versions, 10.5 works really well, 10.6 seems to still need some fixes here and there
[11:04] Vincent.Sylvester @hg.zetaworlds.com:8002: Only had one major crash I still haven't found the source of really
[11:05] Andrew Hellershanks: I haven't had mariadb crash on me. I must not be pushing it hard enough. :)
[11:05] Ubit Umarov: mysql on my box does crash on almost windows restart :)
[11:05] Ubit Umarov: damages a db no idea why
[11:05] Gavin.Hird @grid.xmir.org:8002: The MariaDB Foundation is pleased to announce the availability of MariaDB 10.6.0, the first alpha release in the new MariaDB 10.6 development series.
[11:06] Gavin.Hird @grid.xmir.org:8002: so
[11:06] Gavin.Hird @grid.xmir.org:8002: not fit for purpose in any kind of production setting
[11:06] Ubit Umarov: yeah the one that forced me to release a conenctor fix
[11:06] Gavin.Hird @grid.xmir.org:8002: don't even promote it at all
[11:06] Ubit Umarov: bc he does not know what utf8mb3 is
[11:07] Ubit Umarov: well the fix was not mine.. i just used a patch on mantis :)
[11:07] Vincent.Sylvester @hg.zetaworlds.com:8002: I mean this far into dev cycle the chances an alpha that is just an incremental update to have significant bugs is somewhat slimmer than their first versions, but yeah wouldn't want to use that yet, it seems fine, but who knows what's lurking
[11:08] Ubit Umarov: well no one should use alpha things
[11:08] Ubit Umarov: except that opensim thing..  ethernal alpha
[11:08] Vincent.Sylvester @hg.zetaworlds.com:8002: Was about to say lol
[11:09] Ubit Umarov: it should be part of the name even
[11:09] Gavin.Hird @grid.xmir.org:8002: ok for experimentation, but not for saving anything you want to keep for yourself or your users
[11:09] Ubit Umarov: OpenSimulatorAlpha
[11:09] Ubit Umarov: sounds nice no?
[11:09] Andrew Hellershanks: :)
[11:09] Gavin.Hird @grid.xmir.org:8002: Is that lice DEC  Alpha?
[11:09] Gavin.Hird @grid.xmir.org:8002: like*
[11:09] Ubit Umarov: well hope better than that one :)
[11:09] Gavin.Hird @grid.xmir.org:8002: :-))
[11:10] Vincent.Sylvester @hg.zetaworlds.com:8002: Had simulators run for a year without needing a restart or crashing so if that's an alpha then what will the beta be like, solve world hunger or something, setting the bar quite high he
[11:11] Ubit Umarov: only problem of opensim is users
[11:11] Ubit Umarov: no users no problems.. no crashes
[11:11] Ubit Umarov: :)
[11:11] Vincent.Sylvester @hg.zetaworlds.com:8002: If you mis-use any software you'll have a bad time, just look at Windows
[11:11] Ubit Umarov: well all software is perfect without users
[11:11] Gavin.Hird @grid.xmir.org:8002: so your saying opensim  nver crahs even without users?
[11:12] Ubit Umarov: shhh ignore those details gavin :p
[11:12] Gavin.Hird @grid.xmir.org:8002: :-)
[11:12] Ubit Umarov: who cares if opensim crashes if there are no users to see ?
[11:12] Vincent.Sylvester @hg.zetaworlds.com:8002: That's the tree in the woods thing again
[11:12] Gavin.Hird @grid.xmir.org:8002: it is
[11:12] Andrew Hellershanks: Sounds like one of those "if a tree falls in the forest..." questions.
[11:12] Ubit Umarov: well and we arrived to shrodinger's cat kidda issues :)
[11:13] Ubit Umarov: better change topic :)
[11:13] Vincent.Sylvester @hg.zetaworlds.com:8002: If you have them as service to auto restart and then just not monitor them everything is working fine no matter what
[11:13] Ubit Umarov: so what is new on opensim?
[11:13] Andrew Hellershanks: The news about OpenSim this week is that there is no news.
[11:14] Ubit Umarov: no code changes this last week :(
[11:14] Andrew Hellershanks: None.
[11:14] Ubit Umarov: give time for ppl to crash with previus ones :)
[11:14] Andrew Hellershanks: That must be it. :)
[11:14] Vincent.Sylvester @hg.zetaworlds.com:8002: I wanted to work a bit on a patch that overhauls the way access control is done, have most of it done and working, just not polished
[11:15] Gavin.Hird @grid.xmir.org:8002: that's the spirit!
[11:15] Vincent.Sylvester @hg.zetaworlds.com:8002: The idea is to make Access control section central for hypergrid and local access control rather than have duplicate sections in the config file for both
[11:15] Vincent.Sylvester @hg.zetaworlds.com:8002: Problem is that does mean less granular control over it
[11:16] Vincent.Sylvester @hg.zetaworlds.com:8002: So still working that conundrum out
[11:16] Ubit Umarov: ( did you test that remote parcels thing gavin? )
[11:16] Gavin.Hird @grid.xmir.org:8002: the one I fixed my code for?
[11:17] Ubit Umarov: ( bad region handle on 2 points )
[11:17] Andrew Hellershanks: Vincent, you can discuss the issue via the mailing list or the IRC dev channel if you want to bounce ideas off other people.
[11:17] Gavin.Hird @grid.xmir.org:8002: only on 1 point in my code
[11:17] Gavin.Hird @grid.xmir.org:8002: the rest was some FS mess
[11:17] Ubit Umarov: think you told you had fix on onr yhr places
[11:17] Ubit Umarov: the other is not on yr code?
[11:17] Gavin.Hird @grid.xmir.org:8002: no
[11:18] Ubit Umarov: ok
[11:18] Ubit Umarov: no idea why that was using region center
[11:18] Ubit Umarov: region center is meaning less are region identifier
[11:18] Ubit Umarov: less as region identifier, i meant
[11:19] Vincent.Sylvester @hg.zetaworlds.com:8002: Need to see if there is a potential for bad overlapping in mac bans first, but need to revise that anyhow given how there are issues with where mac is taken from anyways
[11:20] Ubit Umarov: that code was sending invalid (handle, offset) pairs that did not pointed to the required global point
[11:20] Vincent.Sylvester @hg.zetaworlds.com:8002: Ideally comprehensive references to other known user data to validate you have the correct person, but code is never as smart as a human in that regard
[11:20] Ubit Umarov: opensim now should work fine with valid pairs
[11:20] Gavin.Hird @grid.xmir.org:8002: as far as I could find ti wa specific to FS, so no idea why it was like that
[11:21] Ubit Umarov: ofc when possible the they should be relative to region reference corner
[11:21] Ubit Umarov: that makes life easier
[11:22] Ubit Umarov: but not sure all protocol paths allow for offsets > 256m..  guess all do
[11:22] Ubit Umarov: if not, another closer handler must be used
[11:23] Ubit Umarov: just both need to ne coerent using same grid reference point
[11:23] Ubit Umarov: well ok i shutup .. next?
[11:23] Vincent.Sylvester @hg.zetaworlds.com:8002: I did encounter some noteworthy exchange this week that there are still folks that mix physics and meshing up to using ubode with Meshmerizer and XEngine, I was under the impression that was no longer possible
[11:23] Ubit Umarov: yeah even a distro does that
[11:24] Ubit Umarov: calling it "hybird physics"
[11:24] Ubit Umarov: hybrid
[11:24] Ubit Umarov: insane..  but well
[11:24] Ubit Umarov: im sure many still try to use square wheels on cars
[11:24] Vincent.Sylvester @hg.zetaworlds.com:8002: In master is there something to prevent mixing of physics and meshing though?
[11:25] Ubit Umarov: no
[11:25] Ubit Umarov: code assumes ppl have brains
[11:25] Ubit Umarov: even more.. brains that work a bit :p
[11:26] Vincent.Sylvester @hg.zetaworlds.com:8002: heh a bold assumption to make, well most do see the connection given the names I guess, just appears some still don't
[11:26] Ubit Umarov: i do remember lkalif when i did add ubodemeshmerizer
[11:26] Ubit Umarov: he did ask me to merge them
[11:26] Ubit Umarov: a promisse i still have to do
[11:27] Ubit Umarov: just it is a bit of a pain
[11:27] Vincent.Sylvester @hg.zetaworlds.com:8002: I can imagine
[11:27] Grid: George Equus gave you Script error Patrol Bot.
[11:27] Ubit Umarov: guess bullet and old ode could use the cleaner meshes for the tortured prims also..  bit well
[11:28] Ubit Umarov: they don't seem to complain that much with current ones they get
[11:29] Vincent.Sylvester @hg.zetaworlds.com:8002: Bullet still has like two three edge cases it works well for, but ubode as a whole package I have only heard praise of and YEngine in conjunction just a pleasure to work with given the proximity to LSL standard which makes a lot of things easier to debug
[11:30] Vincent.Sylvester @hg.zetaworlds.com:8002: "Issues" there almost exclusively related to bad mesh or scripting which is mostly easy to fix if you know how
[11:30] Ubit Umarov: well non standard lsl things is also a Yengine option :)
[11:30] Andrew Hellershanks: That makes me think that a lot of bad code is likely to stick around if it (now) works in OS.
[11:30] Ubit Umarov: for now just a few
[11:31] Ubit Umarov: and the bad meshes ubOde will only do as convex
[11:31] Vincent.Sylvester @hg.zetaworlds.com:8002: Bad code is a given Andrew it's like the sun being a huge ball of fire
[11:31] Ubit Umarov: tons of them, people just refuse to replace
[11:31] Gavin.Hird @grid.xmir.org:8002: refurse...
[11:31] Vincent.Sylvester @hg.zetaworlds.com:8002: A lot don't know how or the original creators refuse to for some reason
[11:32] Andrew Hellershanks: Vincent, true but if a bad script broke in OS it is likely to get fixed and it will wind up running better than the original.
[11:32] Gavin.Hird @grid.xmir.org:8002: usually it is a case of not being able to
[11:32] Ubit Umarov: well many original creators are gone
[11:32] Ubit Umarov: but sure there are new ones..
[11:33] Vincent.Sylvester @hg.zetaworlds.com:8002: Comes down to content ownership and such legal rights jumbo, but in the end there are always ways to fix things thankfully
[11:33] Ubit Umarov: a few ppl even likes to fix the houses collisions :)
[11:33] Ubit Umarov: so the bad mesh is only used on visual
[11:33] Ubit Umarov: and needed collisions are done by invisible prims
[11:34] Ubit Umarov: boxes hopefuly
[11:34] Gavin.Hird @grid.xmir.org:8002: problem is invisible prims adds to rendering overhad
[11:34] Grid: Notecard updated
[11:34] Ubit Umarov: that must be a bad render
[11:34] Gavin.Hird @grid.xmir.org:8002: the renderer sees averything and only removes in final passes
[11:35] Ubit Umarov: that does not take them out fast
[11:35] Grid: Notecard updated
[11:35] Ubit Umarov: it should take them out fast of main transparency is 100%
[11:36] Ubit Umarov: i do that on warp3d map, ofc
[11:36] Gavin.Hird @grid.xmir.org:8002: there are lots of other cases where 100% transparency are used that should not be taken out so it is not that easy
[11:36] Vincent.Sylvester @hg.zetaworlds.com:8002: Is that based on transparency value? You can set the texture to transparent as well leaving transparency value at 0
[11:36] Ubit Umarov: no not texture
[11:37] Ubit Umarov: on texture the viewer must to it all
[11:37] Gavin.Hird @grid.xmir.org:8002: for all the renderer know it can be a window prim spanning multiple other prims
[11:37] Vincent.Sylvester @hg.zetaworlds.com:8002: Interesting, I used to do both, set texture transparent and transparency to 100
[11:37] Ubit Umarov: while base traspancy it is just a number
[11:38] Ubit Umarov: easy to see it means 100% and skup the entire face fast
[11:38] Ubit Umarov: skip
[11:38] Ubit Umarov: well i do that on warp3d
[11:38] Ubit Umarov: so should viewers
[11:39] Ubit Umarov: i did consider not sending prims with full transparence to users witohut mod rigths
[11:39] Joe Magarac: Base transparency > 0 disables ambient lighting on the object, at least in Firestorm.
[11:39] Ubit Umarov: but did gave up.. would possible breal other things
[11:40] Ubit Umarov: like prim number
[11:40] Ubit Umarov: i mean linknumber
[11:40] Ubit Umarov: viewers do that by counting prims
[11:40] Ubit Umarov: insane wrong
[11:41] Ubit Umarov: but they always did that
[11:41] Ubit Umarov: assume udp does have a packet order
[11:41] Ubit Umarov: bahh
[11:41] Gavin.Hird @grid.xmir.org:8002: be a good boy and send what the viewer wants you to send
[11:42] Ubit Umarov: yeah so i did gave up that speedup
[11:42] Ubit Umarov: but even so..  viewers will do a mess on linknumber if there is a little error rate on udp
[11:43] Joe Magarac: Ah. I'd wondered how that worked. Is there any effort sim-side to send all the ObjectUpdate messages in one packet?
[11:43] Ubit Umarov: seems several lindens that did work on lludp had no idea about what lludp is
[11:44] Ubit Umarov: then after making a mess out of it, decided to move to insane heavy http
[11:44] Ubit Umarov: and use http as a shor messages protocol
[11:44] Ubit Umarov: a insane nonsense
[11:44] Ubit Umarov: we are paying for today
[11:45] Gavin.Hird @grid.xmir.org:8002: the "cloud" won't allow them to use udp
[11:45] Ubit Umarov: you are wrong
[11:45] Ubit Umarov: http3.0 is udp
[11:45] Gavin.Hird @grid.xmir.org:8002: that is what they calim
[11:45] Gavin.Hird @grid.xmir.org:8002: http 3.0 is not even production ready
[11:45] Ubit Umarov: well you maybe right.. they may had not update
[11:45] Ubit Umarov: ofc ti is
[11:45] Ubit Umarov: all google sites use it
[11:45] Gavin.Hird @grid.xmir.org:8002: so it was not an option when management decided they need to move to cloud
[11:46] Gavin.Hird @grid.xmir.org:8002: google is a mess in its own
[11:46] Ubit Umarov: not sure hwo that will not kill vidio and audion
[11:46] Gavin.Hird @grid.xmir.org:8002: ass I said last time: Friends don't let friends use Google
[11:46] Ubit Umarov: bc those do depend on udp having higher priority than tcp
[11:46] Vincent.Sylvester @hg.zetaworlds.com:8002: A cloud move almost always comes from middle management thinking they can reduce costs when the added complexity and potential for failure requiring redundant systems comes out to making the cloud provider rich, not you :)
[11:47] Ubit Umarov: now with all http contents on udp..  well  lets see
[11:47] Gavin.Hird @grid.xmir.org:8002: does LL have middle managment?
[11:47] Vincent.Sylvester @hg.zetaworlds.com:8002: I thought LL was only middle management
[11:47] Gavin.Hird @grid.xmir.org:8002: :-)
[11:47] Ubit Umarov: but that just to tell that LL should had improved lludp fixing its problems
[11:47] Gavin.Hird @grid.xmir.org:8002: at times the entire company seems more like a seminar than anything esle
[11:48] Vincent.Sylvester @hg.zetaworlds.com:8002: I have seen more CEO potential in some toilet brushes quite frankly
[11:48] Joe Magarac: You don't have to use HTTP 3.0, even with Google sites. It will fall back to 2.0 and TCP.  Mostly Chrome talks to Google in HTTP 3.0. Not much else does.
[11:48] Ubit Umarov: ofc it is possivel to make good message protocol on top of udp, QUICK is a full  working stack
[11:48] Ubit Umarov: edge also :p
[11:49] Joe Magarac: Edge has Chrome inside now.
[11:49] Ubit Umarov: yeap
[11:49] Gavin.Hird @grid.xmir.org:8002: Edge is a skin that reports your browsing also to Microsoft
[11:49] Ubit Umarov: but my point is just to say how bad ll was just moving to silly and heavy http 1.0
[11:50] Ubit Umarov: when it was possible to fix lludp
[11:50] Vincent.Sylvester @hg.zetaworlds.com:8002: It's called that cause it's mean to make you want to jump off the Edge of a cliff I swear
[11:50] Gavin.Hird @grid.xmir.org:8002: :-)
[11:50] Ubit Umarov: ofc somethings where ok to move to http
[11:50] Ubit Umarov: request for large things are better on http
[11:50] Gavin.Hird @grid.xmir.org:8002: In Win 11 it is apparently more or less impossible to get rid of
[11:51] Vincent.Sylvester @hg.zetaworlds.com:8002: I was happy with Win7, Win10 annoys me daily, I don't even what to think what 11 will bring, I value what remains of my sanity
[11:51] Ubit Umarov: i do remember how ms killed all browsers forcing the use of IE
[11:52] Ubit Umarov: i was a netscape (?) user
[11:52] Gavin.Hird @grid.xmir.org:8002: me too
[11:52] Gavin.Hird @grid.xmir.org:8002: back in the day...
[11:52] Ubit Umarov: yeah
[11:52] Jamie.Jordan @grid.kitely.com:8002: mr too
[11:52] Jamie.Jordan @grid.kitely.com:8002: me
[11:52] Jagga Meredith: Thanks for all the help with my tp-code issues.  Part of problem was, in fact, that I was running on old code with bugs.  Do we have an "official" 9.2 I can point people to? ie. not beta/dev?
[11:52] Andrew Hellershanks: It was my first browser, IIRC.
[11:52] Ubit Umarov: all opensim is Beta :p
[11:53] Andrew Hellershanks: Ubit, don't you mean  Alpha?
[11:53] Gavin.Hird @grid.xmir.org:8002: so it would be 0.9.2
[11:53] Ubit Umarov: no ALpha will be last stage :)
[11:53] Ubit Umarov: well counting backwards
[11:53] Gavin.Hird @grid.xmir.org:8002: RC alpha
[11:54] Ubit Umarov: about dev..  i should release a non dev version
[11:54] Jagga Meredith: any particular rev?
[11:54] Ubit Umarov: abotu to do it for a year now :)
[11:54] Ubit Umarov: then keep changing code...
[11:55] Jagga Meredith: yeah, people are waiting for it
[11:55] Gavin.Hird @grid.xmir.org:8002: Opensim 2022
[11:55] Andrew Hellershanks hands Ubit a round TUIT.
[11:55] Jagga Meredith: heh
[11:55] Ubit Umarov: osgrif does releases
[11:55] Gavin.Hird @grid.xmir.org:8002: Like Visual Studio 2022 or Office 2022
[11:55] Ubit Umarov: last is from last 16th
[11:55] Ubit Umarov: but osgrid always is on opensim dev versions
[11:56] Ubit Umarov: other grids are on 0.9.2
[11:56] Jagga Meredith: want to run on another grid
[11:56] Gavin.Hird @grid.xmir.org:8002: isn't osgrid a test grid?
[11:56] Andrew Hellershanks: Gavin, it is.
[11:56] Ubit Umarov: is a mix test/operational
[11:56] Gavin.Hird @grid.xmir.org:8002: so
[11:56] Vincent.Sylvester @hg.zetaworlds.com:8002: Running master dev in production for years and have no complaints really, what bugs there are get fixed quite fast these days
[11:57] Vincent.Sylvester @hg.zetaworlds.com:8002: What's broken has either become a feature now or no one has found it yet
[11:57] Ubit Umarov: :)
[11:58] Vincent.Sylvester @hg.zetaworlds.com:8002: The parcel fix or recent has been in the code how long and only recently even showed up on my radar
[11:58] Vincent.Sylvester @hg.zetaworlds.com:8002: Sometimes it takes time for edge cases to show their ugly heads
[11:59] Ubit Umarov: at osgrid the regions that take most fresh bugs are this, mine and lbsa
[11:59] Vincent.Sylvester @hg.zetaworlds.com:8002: There is still a bug with prim scaling, but I have been unable to pinpoint the cause for years now, beginning to think it's base udp stuff on scale update that just gets lost somewhere, finding that is a needle in a haystack
[11:59] Ubit Umarov: this bill did fully automatic
[12:00] Andrew Hellershanks: I have heard that the recent osgrid crash has caused a number of problems for some people. Some missing prims, and some scripted objects that previously worked, don't.
[12:00] Joe Magarac: Thanks all.
[12:00] Ubit Umarov: sees a new commit, compiles and fires here
[12:00] Andrew Hellershanks: Bye, Joe.
[12:00] Ubit Umarov: well those problems need to be reported on osgrid foruns
[12:01] Ubit Umarov: i don't see how scripts that worked stopped, bc assets services down for a day
[12:01] Vincent.Sylvester @hg.zetaworlds.com:8002: Notecards perhaps
[12:01] Ubit Umarov: soem ofc may need a reset
[12:01] George Equus: I need to go. Thanks all.  Bye.
[12:02] Ubit Umarov: cya
[12:02] Andrew Hellershanks: ok, George. tc. See you again some time.
[12:02] Vincent.Sylvester @hg.zetaworlds.com:8002: Mostly just an annoyance how osnotecard just makes a mess making a new asset for each time it is called, but eh cleaning that is near impossible
[12:02] Gavin.Hird @grid.xmir.org:8002: often crashes like that requires full rebuild of db tables where inconsistencies are discarded. Inconsistencies that will show up in world
[12:03] Ubit Umarov: it does not make a new asset
[12:03] Ubit Umarov: hmm does it? :
[12:03] Vincent.Sylvester @hg.zetaworlds.com:8002: Kinda well if data is same it doesn't I think
[12:03] Ubit Umarov: it either does or does not
[12:03] Vincent.Sylvester @hg.zetaworlds.com:8002: In about 3million assets I found 700000 notecards
[12:03] Ubit Umarov: can't have middle thing
[12:03] Jagga Meredith: my region runningh fine without reboot.  saw a bunch of asset updates
[12:03] Vincent.Sylvester @hg.zetaworlds.com:8002: all descript "script generated"
[12:04] Ubit Umarov: npcs are a case ofc
[12:04] Ubit Umarov: clone appearance to a notecard
[12:05] Vincent.Sylvester @hg.zetaworlds.com:8002: You can clean that data out if you can see if the notecard is still in primitems or inventory, but that is a pretty big sql query and will run slow given all the databases you need to check to make sure
[12:05] Vincent.Sylvester @hg.zetaworlds.com:8002: Possible, but not practical
[12:05] Ubit Umarov: new script does irritate me more
[12:05] Ubit Umarov: every one is a new asset :)
[12:06] Ubit Umarov: worse any we try to compile with bugs is asset
[12:06] Andrew Hellershanks: I often wonder when saving an updated script or notecard whether is updates the asset or creates a new one every time.
[12:06] Ubit Umarov: we should only allow at least the ones that compile
[12:06] Ubit Umarov: creates a new one
[12:06] Ubit Umarov: there are no asset updates
[12:06] Ubit Umarov: even viewers dictate a new one
[12:06] Gavin.Hird @grid.xmir.org:8002: so you have never had the need to save a WIP that still did not compile?
[12:07] Ubit Umarov: yeah a problem gaiv
[12:07] Andrew Hellershanks: Ubit, that seems wasteful to me. Lots of script files and notecard assets that are essentially dead cluttering up the db table.
[12:07] Ubit Umarov: just a optin issue
[12:08] Ubit Umarov: tons and tons of garbage added every day to disks
[12:08] Vincent.Sylvester @hg.zetaworlds.com:8002: Well the problem is figuring out what's useful to keep and what not is not a clear case
[12:08] Ubit Umarov: we are alive bc disks keep getting bigger
[12:08] Ubit Umarov: at same cost basicly
[12:08] Vincent.Sylvester @hg.zetaworlds.com:8002: SSDs too otherwise all that data served in that short a time
[12:09] Vincent.Sylvester @hg.zetaworlds.com:8002: OpenSim on HDD can get slow fast
[12:09] Ubit Umarov: od of those d's means disk :p
[12:09] Ubit Umarov: one of..
[12:10] Ubit Umarov: oops ssd only has one D that does mean disk
[12:10] Ubit Umarov: details..
[12:10] Gavin.Hird @grid.xmir.org:8002: proper raid of HDD can be as fast as SSD, and will most likely not stall as fast in a busy invironment
[12:10] Ubit Umarov: ans possible still last longer
[12:10] Gavin.Hird @grid.xmir.org:8002: once SSD cache is full ig tanks completely
[12:10] Gavin.Hird @grid.xmir.org:8002: it*
[12:10] Jagga Meredith: are we storing every plain ordinary "box" or is that a pointer?
[12:11] Ubit Umarov: ssd life time is a hidden secret
[12:11] Vincent.Sylvester @hg.zetaworlds.com:8002: I have yet to play with more complex things like ssd caching hdd raids, sounds like more points of failure though
[12:11] Ubit Umarov: all pub about how fast.. not telling that they did die fast..
[12:11] Ubit Umarov: hope they are better now
[12:11] Ubit Umarov: at least im using one
[12:11] Vincent.Sylvester @hg.zetaworlds.com:8002: I had two SSDs fail at the same time, had to replace one then the other to save the data, that was fun
[12:12] Ubit Umarov: but i have viewer caches on a HDD
[12:12] Ubit Umarov: not going to burn the ssd with those
[12:12] Gavin.Hird @grid.xmir.org:8002: the latests thing with SSDF that have started to appear is that manufacturers replace original components with slower and cheaper once the first version has been benchmarked by all sites that does that stuff
[12:12] Gavin.Hird @grid.xmir.org:8002: SSD*
[12:12] Ubit Umarov: yeah
[12:12] Vincent.Sylvester @hg.zetaworlds.com:8002: Sounds like car manufacturers all over
[12:12] Ubit Umarov: but think that is a old thing on all hardware
[12:13] Ubit Umarov: pass reviewers with premium versions
[12:13] Ubit Umarov: then sell crap
[12:13] Andrew Hellershanks: A project I'm working on now stores data on a system that only has an SD memory card. I use a RAM disk to minimize writing to the disk for temporary files.
[12:13] Ubit Umarov: yeah guess all to that
[12:13] Ubit Umarov: its modern ethics
[12:13] Vincent.Sylvester @hg.zetaworlds.com:8002: I can hear Volkswagen heavy breathing in the corner lol
[12:14] Ubit Umarov: yeah hdds are fading out form market
[12:14] Ubit Umarov: from..
[12:14] Gavin.Hird @grid.xmir.org:8002: there are new HDDs that are about as fast as SATA SSD
[12:14] Ubit Umarov: well nvm with pcie 4 seem to just fly
[12:14] Vincent.Sylvester @hg.zetaworlds.com:8002: Yet sas is still stupid expensive, so is tape
[12:14] Ubit Umarov: yeha but who wants sata ?
[12:15] Andrew Hellershanks: I'd like to see SSDs that use that new form of Flash memory that doesn't wear out.
[12:15] Vincent.Sylvester @hg.zetaworlds.com:8002: What wouldn't I give for a tape drive...
[12:15] Ubit Umarov: sata his limited by spec
[12:15] Gavin.Hird @grid.xmir.org:8002: plenty of systems around with it still
[12:15] Ubit Umarov: nvm now does 8GB/s they claim
[12:15] Ubit Umarov: on pcie4.0
[12:16] Ubit Umarov: sata does at most 600MB/s if i remember by spec
[12:16] Gavin.Hird @grid.xmir.org:8002: but you can still raid it for much faster total
[12:16] Ubit Umarov: well so you can on nvm
[12:16] Andrew Hellershanks: Vincent, be careful what you wish for. I used to use 4mm DAT tapes to do backups. The tapes were fine. The SCSI based DAT drives all died on me. Two died in different ways so I combined them in to a new working drive until it died. A third drive later died.
[12:16] Ubit Umarov: on a EPIC cpu :)
[12:16] Ubit Umarov: or with tons of pcie  lanes
[12:17] Vincent.Sylvester @hg.zetaworlds.com:8002: All fancy new hardware and china just shut down plants
[12:17] Vincent.Sylvester @hg.zetaworlds.com:8002: We'll be waiting some time
[12:17] Ubit Umarov: insane how market is now
[12:17] Andrew Hellershanks: I'd like to get a working DAT drive again to see if I can recover the data on those tapes to see what was on them.
[12:17] Ubit Umarov: impossivle to get the most basic electronic components
[12:18] Vincent.Sylvester @hg.zetaworlds.com:8002: Monitors started go up in price now as well
[12:18] Ubit Umarov: covod excuse
[12:18] Andrew Hellershanks: I've been holding off ordering parts and PCBs while there is an issue with transporting goods.
[12:18] Ubit Umarov: covid
[12:19] Ubit Umarov: covid made a great excuse for all kind od speculation and extra profits
[12:19] Gavin.Hird @grid.xmir.org:8002: I want one of the massive tape libraries I used to sell when i IBM, with automatic migration between disk and tape (both ways)
[12:19] Ubit Umarov: guess also to all the "just in time" supply chains
[12:19] Andrew Hellershanks: We are getting rather of topic and it is past the hour. Any last minute OS related items for today?
[12:20] Gavin.Hird @grid.xmir.org:8002: There was an update to the Win version of my viewer a few days back. basically adding stuff already in the Mac version
[12:21] Andrew Hellershanks: ty for the info, Gavin.
[12:21] Gavin.Hird @grid.xmir.org:8002: yw
[12:21] Andrew Hellershanks: Any thing else?
[12:21] Andrew Hellershanks: Going...
[12:21] Andrew Hellershanks: Going...
[12:21] Andrew Hellershanks: gone.  :)
[12:21] Andrew Hellershanks: That will do it for this week. Thank you all for coming. See you again next week.
Personal tools
General
About This Wiki