Is the Qt client killing the hard drive slowly?

[manicminer]

I'm just asking this because I read it downloads much faster on an SSD. Even much faster on a RAM disk. I have an older machine. I download the chain for 5 days. Does the downloading process or simply operating the client kill the hard drive?

[1714021426]

1714021426

[1714021426]

1714021426

[1714021426]

1714021426

[niko]

I'm just asking this because I read it downloads much faster on an SSD. Even much faster on a RAM disk. I have an older machine. I download the chain for 5 days. Does the downloading process or simply operating the client kill the hard drive?

No. Hard drives are designed to be used as hard drives, which is exactly what you are doing when using the computer with a hard drive. I hope this is clear enough.

[payb.tc]

^ if downloading the blockchain kills your harddrive, your harddrive was probably at the end of it's life anyway.

[Kluge]

I'm just asking this because I read it downloads much faster on an SSD. Even much faster on a RAM disk. I have an older machine. I download the chain for 5 days. Does the downloading process or simply operating the client kill the hard drive?

No. Hard drives are designed to be used as hard drives, which is exactly what you are doing when using the computer with a hard drive. I hope this is clear enough.

Could you explain why hard drives are different from GPUs or gasoline engines which have "health" decreased with excessive use? For example, assume you have two otherwise identical engines -- one has 10k miles on it, the other has 100k miles on it (or km). Why would the engine with less miles on it fetch a higher price, even if they were manufactured in the same year?

Or for GPUs, a GPU used lightly should fetch a higher price than a GPU which mined for the past 3 years. They are otherwise identical, but one has more "wear."

Why would a hard drive be any different?

[niko]

I'm just asking this because I read it downloads much faster on an SSD. Even much faster on a RAM disk. I have an older machine. I download the chain for 5 days. Does the downloading process or simply operating the client kill the hard drive?

No. Hard drives are designed to be used as hard drives, which is exactly what you are doing when using the computer with a hard drive. I hope this is clear enough.

Could you explain why hard drives are different from GPUs or gasoline engines which have "health" decreased with excessive use? For example, assume you have two otherwise identical engines -- one has 10k miles on it, the other has 100k miles on it (or km). Why would the engine with less miles on it fetch a higher price, even if they were manufactured in the same year?

Or for GPUs, a GPU used lightly should fetch a higher price than a GPU which mined for the past 3 years. They are otherwise identical, but one has more "wear."

Why would a hard drive be any different?

There is some normal wear of mechanical components with use, and this will eventually lead to failure. Before this point is reached, however, other causes may and do trigger failures: abuse, mechanical shocks, ESD, thermal stress cycles, dusty or smoky environment, etc.  The probability of most of these does not increase with use. Failure due to thermal stress is in fact more likely in systems subject to on/off cycles (less use!) than those that are constantly turned on.

Unlike GPUs or HDDs, internal combustion engines are mainly mechanical devices, with rather violent vibrations and stress in normal use, so wear and tear may in fact be a dominant mechanism of failure.

[Kluge]

I'm just asking this because I read it downloads much faster on an SSD. Even much faster on a RAM disk. I have an older machine. I download the chain for 5 days. Does the downloading process or simply operating the client kill the hard drive?

No. Hard drives are designed to be used as hard drives, which is exactly what you are doing when using the computer with a hard drive. I hope this is clear enough.

Could you explain why hard drives are different from GPUs or gasoline engines which have "health" decreased with excessive use? For example, assume you have two otherwise identical engines -- one has 10k miles on it, the other has 100k miles on it (or km). Why would the engine with less miles on it fetch a higher price, even if they were manufactured in the same year?

Or for GPUs, a GPU used lightly should fetch a higher price than a GPU which mined for the past 3 years. They are otherwise identical, but one has more "wear."

Why would a hard drive be any different?

There is some normal wear of mechanical components with use, and this will eventually lead to failure. Before this point is reached, however, other causes may and do trigger failures: abuse, mechanical shocks, ESD, thermal stress cycles, dusty or smoky environment, etc.  The probability of most of these does not increase with use. Failure due to thermal stress is in fact more likely in systems subject to on/off cycles (less use!) than those that are constantly turned on.

Unlike GPUs or HDDs, internal combustion engines are mainly mechanical devices, with rather violent vibrations and stress in normal use, so wear and tear may in fact be a dominant mechanism of failure.

Thanks!

[manicminer]

Clear enough!

[Pieter Wuille]

Current versions of the reference client indeed causes excessive disk I/O. The next version (0.8, not yet released) should improve upon this significantly (it uses a more modern database engine, and a new database layout).

[Meni Rosenfeld]

I'm just asking this because I read it downloads much faster on an SSD. Even much faster on a RAM disk. I have an older machine. I download the chain for 5 days. Does the downloading process or simply operating the client kill the hard drive?

No. Hard drives are designed to be used as hard drives, which is exactly what you are doing when using the computer with a hard drive. I hope this is clear enough.

Could you explain why hard drives are different from GPUs or gasoline engines which have "health" decreased with excessive use? For example, assume you have two otherwise identical engines -- one has 10k miles on it, the other has 100k miles on it (or km). Why would the engine with less miles on it fetch a higher price, even if they were manufactured in the same year?

Or for GPUs, a GPU used lightly should fetch a higher price than a GPU which mined for the past 3 years. They are otherwise identical, but one has more "wear."

Why would a hard drive be any different?

There is some normal wear of mechanical components with use, and this will eventually lead to failure. Before this point is reached, however, other causes may and do trigger failures: abuse, mechanical shocks, ESD, thermal stress cycles, dusty or smoky environment, etc.  The probability of most of these does not increase with use. Failure due to thermal stress is in fact more likely in systems subject to on/off cycles (less use!) than those that are constantly turned on.

Unlike GPUs or HDDs, internal combustion engines are mainly mechanical devices, with rather violent vibrations and stress in normal use, so wear and tear may in fact be a dominant mechanism of failure.

Consumer hard drives are designed to be used as hard drives by consumers. Desktop computers, unlike servers, don't usually access the disks 24/7, so these drives are not designed to withstand such use (you have server-grade drives for that). It will create wear & tear which will shorten the expected component life.

Whether this effect is significant with current usage pattern of bitcoin-qt is not something I can readily answer. It also depends greatly on the specific drive. Some drives park their heads when not actively seeking to spare the wear&tear of the heads hovering close to the platters. Constant usage can interfere with this endurance feature in various ways; it can prevent the heads from ever being parked, or worse, it can create repeated park/unpark operations, for which the drive is only rated for a certain number.

With SSDs there is the well-known write endurance problem, and excessive random write operations can degrade the drive's write performance.

[manicminer]

Current versions of the reference client indeed causes excessive disk I/O.

Only when downloading the database or for normal use as well?

The next version (0.8, not yet released) should improve upon this significantly (it uses a more modern database engine, and a new database layout).

Does it mean a new database download? Cool, when it will be released?

[grue]

The next version (0.8, not yet released) should improve upon this significantly (it uses a more modern database engine, and a new database layout).

Does it mean a new database download? Cool, when it will be released?

not anytime soon because sipa wants to be sure that the new database is stable. you can help test it too: https://bitcointalk.org/index.php?topic=129861.0

[manicminer]

you can help test it too: https://bitcointalk.org/index.php?topic=129861.0

Disclaimer: don't use these for any serious stuff - I am not responsible for destroyed wallets or exploded CPU's.

Well.

[Diapolo]

you can help test it too: https://bitcointalk.org/index.php?topic=129861.0

Disclaimer: don't use these for any serious stuff - I am not responsible for destroyed wallets or exploded CPU's.

Well.

Use it on testnet and help creating some feedback for sipa and the other devs .
Btw.: I think this thread should not be called is the Qt client, but better "the reference client", as bitcoind behaves just the same afaik.

Dia

[Gyrsur]

Current versions of the reference client indeed causes excessive disk I/O. The next version (0.8, not yet released) should improve upon this significantly (it uses a more modern database engine, and a new database layout).

sounds very good, the excessive disk usage of the 0.7 release is terrible on older hardware.

[manicminer]

Disclaimer: don't use these for any serious stuff - I am not responsible for destroyed wallets or exploded CPU's.

Well.

Use it on testnet and help creating some feedback for sipa and the other devs

It won't destroy the wallet for sure but how about the exploded CPU? Can you guarantee with testnet against it? Wondering.

[b!z]

Disclaimer: don't use these for any serious stuff - I am not responsible for destroyed wallets or exploded CPU's.

Well.

Use it on testnet and help creating some feedback for sipa and the other devs

It won't destroy the wallet for sure but how about the exploded CPU? Can you guarantee with testnet against it? Wondering.

Hopefully nothing will blow up.

[niko]

Disclaimer: don't use these for any serious stuff - I am not responsible for destroyed wallets or exploded CPU's.

Well.

Use it on testnet and help creating some feedback for sipa and the other devs

It won't destroy the wallet for sure but how about the exploded CPU? Can you guarantee with testnet against it? Wondering.

Hopefully nothing will blow up.

Hopefully a CPU will blow up and save us from manicminer. In the meantime, there's the "ignore" button.

[manicminer]

Hopefully a CPU will blow up and save us from manicminer. In the meantime, there's the "ignore" button.

?