Popular Content

Showing content with the highest reputation since 02/20/19 in Posts

  1. 1 point
    Please don't judge... and its better just to nod and say, oooohkay. But what was the fastest Hard Drive in 1999? It could have come out as late as the end of '99, but if it released in Jan 1 of '00, it is off the table. I made it to the legacy hard drive benchmark comparison, but it wasn't able to show many any results. I have the Seagate Cheetah 18LP, the Fujitsu MAG3182LP, and the Atlas 10k. I don't think those products were refreshed until the first quarter of '00. I don't know what a user would have chosen for an in home experience back then if money wasn't an option. For some reason I don't remember much about the Cheetah 18LP. I want to say, most of the attention for in home use during that time was focused on the Fujitsu MAG, and the Atlas 10k. I also don't know what was hot in the IDE space at the time.
  2. 1 point
    No need for rewrite, controller does it on it´s own, downgrade to sun fw 12.12.0-0079 let it fail and the load latest fw, exploting a known fw bug it rearms battery capacity as old fw had a bug reporting bad battery capacity ;)
  3. 1 point
    Update: I got confirmed that the test was done in "write through" mode. So this means that, unless anyone would prove otherwise, indeed its RAID 5 and 6 write performance for SSD RAID:s indeed is 25MB/sec per disk set. That is complete junk! I wouldn't even touch that with rubber gloves! So seems anyone who wants data redundancy needs to implement that in software, as hardware simply not provides that, at least not with any performance whatsoever. Incredible, this is 2016, I would really have thought that we were past the 25MB/sec mark on top-of-the-line hardware. Anyhow, thanks for the reality check! Mikael --- Hi everyone - here is an ultra important question: The RAID 5 and RAID 6 benchmark figures are ultra-bad - the write speed has a 98% overhead (6k IOPS = 25MB/sec), and the read has an 80% overhead (130k IOPS = 500MB/sec), for the set of 8 drives. Like, really, so ultra catastrophic that I wouldn't even touch that card - and neither its OEM variant "Megaraid SAS3108". This makes me wonder, however, what settings the benchmark was done with: The recommended settings are to set "Write cache" to "Write through" mode, to enable "Direct I/O" mode, to enable "Disk's cache" (that is, the physical disks' own cache), and to disable the "Read cache" altogether. Perhaps absence of running the test with "CacheVault" would be an indirect reason for the crappy numbers - any thoughts on that? Any completing info on the SSD 4K RW IOPS in RAID 5 and RAID 6 with an actually healthy, proper configuration as described here would be much appreciated!! Please post your thoughts & actual benchmark results - thanks!!
  4. 1 point
    In Linux, you can move and delete/unlink an open file. The file is then deleted when it's closed. In Windows, you can only sometimes rename an open file, but you can't delete an open file. I'm wondering, why not?