Re: My new 160GB WD IDE drive

Discussion in 'NZ Computing' started by Lawrence D'Oliveiro, Apr 16, 2008.

  1. In article <>, Nighthawk did
    write:

    > I just added a new WD160 hard drive as a slave ...


    Can you still get 160GB drives? :)
     
    Lawrence D'Oliveiro, Apr 16, 2008
    #1
    1. Advertising

  2. Lawrence D'Oliveiro

    ~misfit~ Guest

    Somewhere on teh intarweb "Richard" typed:
    > Lawrence D'Oliveiro wrote:
    >> In article <>, Nighthawk
    >> did write:
    >>
    >>> I just added a new WD160 hard drive as a slave ...

    >>
    >> Can you still get 160GB drives? :)

    >
    > Only if you pay almost 2 times the per gig cost of a 750 gig drive.


    The good thing about 160GB drives, at least the Seagate ones anyway, are
    that they're single platter and thinner than a normal HDD. That can make a
    lot of difference where cooling/air flow is an issue.

    Actually I believe that 500GB HDD's are the current sweet-spot for price/GB.
    My New Seagate 500 GB (with 32Mb cache) works out at $0.32 per GB. A far cry
    from my first ever HDD greater than a GB, it was something like $450 for
    1.7GB.
    --
    Shaun.

    > Just waiting for the 1tbs to drop to about the same price/gig so I can
    > build a new array here...
     
    ~misfit~, Apr 16, 2008
    #2
    1. Advertising

  3. Lawrence D'Oliveiro

    ~misfit~ Guest

    Somewhere on teh intarweb "thingy" typed:
    > Richard wrote:
    >> Lawrence D'Oliveiro wrote:
    >>> In article <>, Nighthawk
    >>> did write:
    >>>
    >>>> I just added a new WD160 hard drive as a slave ...
    >>>
    >>> Can you still get 160GB drives? :)

    >
    > Probably get 80s and even 40s....silly money for the Gb mind.
    >
    >> Only if you pay almost 2 times the per gig cost of a 750 gig drive.
    >>
    >> Just waiting for the 1tbs to drop to about the same price/gig so I
    >> can build a new array here...

    >
    > yes....1Tb are a bit too much at the moment, the 750s are a steal
    > though...
    > However, we were discussing SAS/SATA disk failures with an iSCSI
    > vendor and the failure rate for SAS is 1.6%, 18+% for 500GbSATA and
    > 27% for 750Gb SATA....which is worrying....this does not mean they
    > will do that badly for you as these are SATAs used in hvy duty
    > enterprise style workloads and not desktop....still it makes me
    > pause....


    Wow! That is worrying. My main machine is on 24/7. Oh well, at least with
    Seagates I get a 5 year warranty and I back up anything that is *really*
    essential to another HDD in another machine (or two). However, that said, I
    really don't want to lose *any* data. That's why I'm keeping it and buying
    ever bigger HDDs. (I just can't afford to 'mirror' all of it.)

    Just got my RA number from Ascent today to return the drive to Ingram Micro
    in Manukau. I might drop it off in person rather than courier. I haven't
    been out of Pukekohe for too long. Still waiting for the replacement first
    though to try to copy everything off first.

    Cheers,
    --
    Shaun.
     
    ~misfit~, Apr 17, 2008
    #3
  4. Lawrence D'Oliveiro

    ~misfit~ Guest

    Somewhere on teh intarweb "Richard" typed:
    > ~misfit~ wrote:
    >> Somewhere on teh intarweb "Richard" typed:
    >>> Lawrence D'Oliveiro wrote:
    >>>> In article <>, Nighthawk
    >>>> did write:
    >>>>
    >>>>> I just added a new WD160 hard drive as a slave ...
    >>>> Can you still get 160GB drives? :)
    >>> Only if you pay almost 2 times the per gig cost of a 750 gig drive.

    >>
    >> The good thing about 160GB drives, at least the Seagate ones anyway,
    >> are that they're single platter and thinner than a normal HDD. That
    >> can make a lot of difference where cooling/air flow is an issue.
    >>
    >> Actually I believe that 500GB HDD's are the current sweet-spot for
    >> price/GB. My New Seagate 500 GB (with 32Mb cache) works out at $0.32
    >> per GB. A far cry from my first ever HDD greater than a GB, it was
    >> something like $450 for 1.7GB.

    >
    > 1c/gig between them, there are 2 500s listed on pricespy and I dont
    > know which is the 32 meg cache one.
    >
    > Funny that they have more cache on them then my first computer had
    > harddrive space. Few years later I recall spending several 100 on a
    > 800 meg drive which I got specifically since I had trouble burning
    > audio so I had to have somewhere to put the files that was defragged
    > enough to be able to write it at 2 speed without underrunning. Those
    > were not the days, waiting 40 mins to burn an audio disc for the car,
    > on a blank worth almost as much as a pressed cd...


    I remember those days well. I paid $4,500 for one of the first "clones", a
    Total Peripherals DX-2 486/66 with a 428MB HDD. I put a CD RW drive in it
    (at close to $1K) and more than 50% of my burns were coasters. That's why,
    about 18 months after I bought it, I put a 1.7GB Maxtor in it. Within a year
    I had a Cyrix Pentium 166 equivalent (that ran at 150MHz) but a few of my
    games refused to install as they didn't recognise the CPU as being a
    Pentium, which was a minimum requirement.

    It was at that stage that I stopped using computer assemblers completely and
    rolled my sleeves up and did everything myself. However, for a while there i
    couldn't get parts much more cheaply than I could get them installed for, it
    was all "trade only". How times change. Now there's a PC parts vendor in
    every little suburb and sometimes two or three per city block.

    > Now when I run out of space on the serrver its cheaper and easier to
    > whack another drive or 2 in there then to burn old stuff off onto dvd
    > and lable it in triplicate since its so unreliable.


    Aye, I'm in the same boat.
    --
    Shaun.
     
    ~misfit~, Apr 17, 2008
    #4
  5. In article <>, thingy did write:

    > If you used Linux I'd suggest bacula as a backup system...


    rsync FTW!
     
    Lawrence D'Oliveiro, Apr 18, 2008
    #5
  6. Lawrence D'Oliveiro

    Enkidu Guest

    Lawrence D'Oliveiro wrote:
    > In article <>, thingy did write:
    >
    >> If you used Linux I'd suggest bacula as a backup system...

    >
    > rsync FTW!
    >

    rsync. Hm, I don't like rsync. I've had to set up rsync to sync two
    multi-terabyte data stores and I've had nothing but problems. Firstly,
    rsync does not like really large directory structures. It runs out of
    memory and gives a cryptic error message. I had to write some code to
    break up the rsync into multiple rync jobs by iterating through
    sub-directories. Also, rsync when it fails tends to fail with no error
    messages. The only way that you can confirm that it has worked is to
    check the size on disk (du --apparent-size). Even with that the byte
    count is often not identical. I tracked it down after a lot of work to a
    difference in size of the source and destination directories, but the
    data was OK.

    Result is, I've got a 500 line script to sync the data stores where I
    hoped that I would have had a one liner and it's cost me one hell of a
    lot of time to set up. It also takes a long time to run because of the
    time necessary to scan all the directories to make sure that rsync has
    correctly done its job!

    rsync sucks. Its only advantage over something like 'cp -a' is that it
    doesn't move data that has not changed.

    Cheers,

    Cliff

    --

    Have you ever noticed that if something is advertised as 'amusing' or
    'hilarious', it usually isn't?
     
    Enkidu, Apr 19, 2008
    #6
  7. In article <48093b3e$>, Enkidu did write:

    > Firstly, rsync does not like really large directory structures. It runs
    > out of memory and gives a cryptic error message. I had to write some code
    > to break up the rsync into multiple rync jobs by iterating through
    > sub-directories.


    Once you get up to several million files it can use a lot of memory. I've
    done the same.

    > Also, rsync when it fails tends to fail with no error messages.


    Never encountered this. I've had it hit broken connections, I/O errors and
    disks filling up. In all cases the error messages were written in the logs
    loud and clear.

    > The only way that you can confirm that it has worked is to
    > check the size on disk (du --apparent-size). Even with that the byte
    > count is often not identical.


    I check it by running it again. If it reports nothing new to do, then I
    conclude that the transfer succeeded.

    > rsync sucks.


    Compared to what?
     
    Lawrence D'Oliveiro, Apr 19, 2008
    #7
  8. Lawrence D'Oliveiro

    Enkidu Guest

    Lawrence D'Oliveiro wrote:
    > In article <48093b3e$>, Enkidu did write:
    >
    >> Firstly, rsync does not like really large directory structures. It
    >> runs out of memory and gives a cryptic error message. I had to
    >> write some code to break up the rsync into multiple rync jobs by
    >> iterating through sub-directories.

    >
    > Once you get up to several million files it can use a lot of memory.
    > I've done the same.
    >
    >> Also, rsync when it fails tends to fail with no error messages.

    >
    > Never encountered this. I've had it hit broken connections, I/O
    > errors and disks filling up. In all cases the error messages were
    > written in the logs loud and clear.
    >

    Well, I have, frequently. One time the NFS mount disappeared into the
    ether in mid copy and it appeared to have worked with only one third of
    the data copied, but the next rsync of the next directory failed. So I
    decided to *cause* the rsync to fail by pulling the plug on the NFS
    connection but this time it errored!
    >
    >> The only way that you can confirm that it has worked is to check
    >> the size on disk (du --apparent-size). Even with that the byte
    >> count is often not identical.

    >
    > I check it by running it again. If it reports nothing new to do, then
    > I conclude that the transfer succeeded.
    >

    Yes, and if it fails silently in the same way in the same place? The
    only safe way is to check the size on disk in both source and destination.
    >
    >> rsync sucks.

    >
    > Compared to what?
    >

    I'm not comparing it to anything. It sucks all by itself. But it is the
    only tool for the job. It's not a tool for beginners or someone who
    wants it to 'just work'. You have to hedge around it with code that
    checks that everything is OK. Otherwise you could be *thinking* that
    everything was OK, when it wasn't.

    Cheers,

    Cliff

    --

    Have you ever noticed that if something is advertised as 'amusing' or
    'hilarious', it usually isn't?
     
    Enkidu, Apr 19, 2008
    #8
  9. In article <4809747e$>, Enkidu did write:

    > Lawrence D'Oliveiro wrote:
    >
    >> In article <48093b3e$>, Enkidu did write:
    >>
    >>> Also, rsync when it fails tends to fail with no error messages.

    >>
    >> Never encountered this. I've had it hit broken connections, I/O
    >> errors and disks filling up. In all cases the error messages were
    >> written in the logs loud and clear.
    >>

    > Well, I have, frequently. One time the NFS mount disappeared into the
    > ether in mid copy and it appeared to have worked with only one third of
    > the data copied, but the next rsync of the next directory failed.


    Why on Earth are you trying to rsync an NFS-mounted volume? Why not do the
    rsync over the network, directly to/from the volume locally on the NFS
    server? It'll also go a lot faster.
     
    Lawrence D'Oliveiro, Apr 19, 2008
    #9
  10. Lawrence D'Oliveiro

    Enkidu Guest

    Lawrence D'Oliveiro wrote:
    > In article <4809747e$>, Enkidu did write:
    >
    >> Lawrence D'Oliveiro wrote:
    >>
    >>> In article <48093b3e$>, Enkidu did write:
    >>>
    >>>> Also, rsync when it fails tends to fail with no error messages.
    >>>>
    >>>>
    >>> Never encountered this. I've had it hit broken connections, I/O
    >>> errors and disks filling up. In all cases the error messages were
    >>> written in the logs loud and clear.
    >>>

    >> Well, I have, frequently. One time the NFS mount disappeared into
    >> the ether in mid copy and it appeared to have worked with only one
    >> third of the data copied, but the next rsync of the next directory
    >> failed.

    >
    > Why on Earth are you trying to rsync an NFS-mounted volume? Why not
    > do the rsync over the network, directly to/from the volume locally on
    > the NFS server? It'll also go a lot faster.
    >

    I do. I used NFS in the early stages of testing. But it was not that
    much slower than ssh.

    Cheers,

    Cliff

    --

    Have you ever noticed that if something is advertised as 'amusing' or
    'hilarious', it usually isn't?
     
    Enkidu, Apr 19, 2008
    #10
  11. In article <4809c4f4$>, Enkidu did write:

    > Lawrence D'Oliveiro wrote:
    >> In article <4809747e$>, Enkidu did write:
    >>
    >>> Lawrence D'Oliveiro wrote:
    >>>
    >>>> In article <48093b3e$>, Enkidu did write:
    >>>>
    >>>>> Also, rsync when it fails tends to fail with no error messages.
    >>>>>
    >>>>>
    >>>> Never encountered this. I've had it hit broken connections, I/O
    >>>> errors and disks filling up. In all cases the error messages were
    >>>> written in the logs loud and clear.
    >>>>
    >>> Well, I have, frequently. One time the NFS mount disappeared into
    >>> the ether in mid copy and it appeared to have worked with only one
    >>> third of the data copied, but the next rsync of the next directory
    >>> failed.

    >>
    >> Why on Earth are you trying to rsync an NFS-mounted volume? Why not
    >> do the rsync over the network, directly to/from the volume locally on
    >> the NFS server? It'll also go a lot faster.
    >>

    > I do. I used NFS in the early stages of testing. But it was not that
    > much slower than ssh.


    Also I think you weren't getting I/O errors when the NFS volume went away
    simply because your NFS client stack wasn't reporting any errors. So how
    was rsync supposed to know something had gone wrong?
     
    Lawrence D'Oliveiro, Apr 19, 2008
    #11
  12. Lawrence D'Oliveiro

    Enkidu Guest

    Lawrence D'Oliveiro wrote:
    > In article <4809c4f4$>, Enkidu did write:
    >
    >> Lawrence D'Oliveiro wrote:
    >>> In article <4809747e$>, Enkidu did write:
    >>>
    >>>> Lawrence D'Oliveiro wrote:
    >>>>
    >>>>> In article <48093b3e$>, Enkidu did
    >>>>> write:
    >>>>>
    >>>>>> Also, rsync when it fails tends to fail with no error
    >>>>>> messages.
    >>>>>>
    >>>>>>
    >>>>> Never encountered this. I've had it hit broken connections,
    >>>>> I/O errors and disks filling up. In all cases the error
    >>>>> messages were written in the logs loud and clear.
    >>>>>
    >>>> Well, I have, frequently. One time the NFS mount disappeared
    >>>> into the ether in mid copy and it appeared to have worked with
    >>>> only one third of the data copied, but the next rsync of the
    >>>> next directory failed.
    >>> Why on Earth are you trying to rsync an NFS-mounted volume? Why
    >>> not do the rsync over the network, directly to/from the volume
    >>> locally on the NFS server? It'll also go a lot faster.
    >>>

    >> I do. I used NFS in the early stages of testing. But it was not
    >> that much slower than ssh.

    >
    > Also I think you weren't getting I/O errors when the NFS volume went
    > away simply because your NFS client stack wasn't reporting any
    > errors. So how was rsync supposed to know something had gone wrong?
    >

    Oh, it knows if a file goes away while it is working. It actual tells
    you "File X has gone away" or similar. That's what I would have
    expected, either at re-read of the directory or at file access time. It
    holds a lot of the file system structure in memory (which is why it has
    problems with large file systems) while it is working. NFS should have
    given at least a "Stale File Handle" error.

    Cheers,

    Cliff

    --

    Have you ever noticed that if something is advertised as 'amusing' or
    'hilarious', it usually isn't?
     
    Enkidu, Apr 19, 2008
    #12
    1. Advertising

Want to reply to this thread or ask your own question?

It takes just 2 minutes to sign up (and it's free!). Just click the sign up button to choose a username and then you can ask your own questions on the forum.
Similar Threads
  1. Silverstrand

    160GB Seagate Momentus 5400.3 2.5" Hard Drive

    Silverstrand, Jan 31, 2006, in forum: Front Page News
    Replies:
    0
    Views:
    715
    Silverstrand
    Jan 31, 2006
  2. Silverstrand
    Replies:
    0
    Views:
    642
    Silverstrand
    Feb 25, 2006
  3. hugh jass

    160Gb hard drive not found on My Computer

    hugh jass, Jan 8, 2005, in forum: Computer Support
    Replies:
    5
    Views:
    1,150
    Conor
    Jan 8, 2005
  4. Blarney Stone

    Maxtor 160GB hard drive question

    Blarney Stone, Feb 14, 2005, in forum: Computer Support
    Replies:
    3
    Views:
    2,029
    Sunny
    Feb 14, 2005
  5. ~misfit~

    Re: My new 160GB WD IDE drive

    ~misfit~, Apr 16, 2008, in forum: NZ Computing
    Replies:
    1
    Views:
    750
    ~misfit~
    Apr 17, 2008
Loading...

Share This Page