Ecogreen vs Western Digital Caviar green

[jim]

Just bought a pair of HD203WIs, very happy with them, and they were even cheaper than the WD EARS drives when I bought them.

[peterb]

+1 for WD, fast quiet, reliable

I have 2 WD ears drives in my main rig (raid1, win7) and also an other EARS with Ubuntu...works perfectly, just investigate if you are planning to use obsolete OSs on EARS drives.

Many say that Samsung drives are just as good, but I won't risk it again for myself. And as I read in reviews the samsung drives shine in synthetic benchmarks but in real world scenarios the WDs have some advantage

anyway good luck

Not really a question of "obsolete" OS, more a question of some of the tools in the distro not fully recognising the drive reports - which means that partitioning has to be done carefully to ensure partitions end on a 4K boundary. It may also be possible to apply the jumper setting required for WinXP which seems to force the drive to 512byte sectors. Ubuntu's distro may have updated fdsk and parted - or you could have been lucky in your partitioning! As I was adding a drive to RAID on production system, I didn't really have the option of tinkering with the partition table on the existing drive.

Hmm. My NAS solution uses OpenFiler + ZFS, and there are spurious reports of problems with these drives due to the 4096 byte block. I'm think I'll do with the Samsungs....

Thanks - you just saved me a lot of time and hassle!

Probably the safest option if you are in a hurry - or don't want the hastle of experimenting.

[jim]

It may also be possible to apply the jumper setting required for WinXP which seems to force the drive to 512byte sectors.

I don't know what you're specifically saying there since you say "seems", but anyway - it doesn't actually force the drive to use smaller sectors - it can't do. The whole benefit of bigger sectors is that less physical disk space is required for the same size partition, so a 1.5TB EARS drive is smaller than a 1.5TB EADS drive, as I understand it. If not, then it's just an EADS with chunk of the drive disabled.

The jumper setting just internally shifts all requests one sector ahead. XP normally installs its partitions 1 sector before the 4kb boundaries, which apparently wrecks performance, so the jumper manually tricks it into installing one sector ahead, which means it falls into line with the start of the next 4kb sectors. So it's not really doing anything clever, and I don't know whether it works with unix based systems - it definitely would not be a good idea to use the jumper in Windows 7 though.

[peterb]

I don't know what you're specifically saying there since you say "seems", but anyway - it doesn't actually force the drive to use smaller sectors - it can't do. The whole benefit of bigger sectors is that less physical disk space is required for the same size partition, so a 1.5TB EARS drive is smaller than a 1.5TB EADS drive, as I understand it. If not, then it's just an EADS with chunk of the drive disabled.

The jumper setting just internally shifts all requests one sector ahead. XP normally installs its partitions 1 sector before the 4kb boundaries, which apparently wrecks performance, so the jumper manually tricks it into installing one sector ahead, which means it falls into line with the start of the next 4kb sectors. So it's not really doing anything clever, and I don't know whether it works with unix based systems - it definitely would not be a good idea to use the jumper in Windows 7 though.

Thank you for clarifying that. The reason I said 'seems' is that it looked as if there was a choice of formatting systems installed on the drive, and the first time the drive was written to, it selected the appropriate one, but I didn't go into it in any great depth as I wasn't using Win XP on it - and having seen the problems other Linux users experienced, I just RMAd the two of them (took them back to Novatech last Sunday (BH Sunday!!) - their CS dept was open!) after getting an RMA authorisation earlier in the week when I ordered the Samsungs. Should get the refund this week. (Mine are the 1.5TB UI variant)

I did see that there is a WD software tool that achieves the same thing as the jumper, but didn't have the time or inclination to download it and faff about!

[jim]

Should get the refund this week. (Mine are the 1.5TB UI variant)

I did see that there is a WD software tool that achieves the same thing as the jumper, but didn't have the time or inclination to download it and faff about!

Yeah, I've got one of the 1.5TB UI drives, my server likes Samsungs

It's not worth it really in my view, mucking around with these new format drives. If they were exceptionally cheaper than the competition, or exceptionally faster, then I'd think fair enough, but they just aren't. Even if you buy it for a Windows 7 machine now, I'd be pretty hacked off if in 5 months' time I can't put it in my server.

I'm sure in a few years time it'll be a different situation, when all drives are like that, but for now it seems fairly irrelevant.

[peterb]

I have also noticed that the Samsungs are running about 5C cooler than the WD (according to the S.M.A.R.T report.

[jim]

Yeah, I noticed that as well with my drives... I'd assumed it was an error, because originally I was comparing 7200rpm Samsung F1 drives with a 500GB Western Digital Caviar Green drive, at around 5900rpm. Yet the Samsungs were always 3-4 degrees cooler.

Surely it must be to do with the system of measurement then? As I say, I'd assumed it was an anomaly or a faulty sensor on the WD, because it seemed impossible, but if you're seeing it as well... weird.

[scaryjim]

if it's consistent across all WD and Sammy drives, then I'd assume it's a different point of measurement which means the WD will always report a higher temp: 3 - 5 degrees is a pretty significant difference otherwise...

[peterb]

There are extra paramters on the Samsung - airflow temp (which is some internal measurement), and it also records max and min temperatures ever reached (I'm using smartctl and smartd under Linux) I wondered if it because the Samsungs are drawing slightly less power. Both the Sammy and the WD have a case fan in front of them but the external airflow across thenm is pretty similar. The other Samsung is mounted slightly differently, with slightly poorer airflow across the chassis but is within a degree of the other - and about 5 below the WD.

WD

193 Load_Cycle_Count 0x0032 179 179 000 Old_age Always - 64066
194 Temperature_Celsius 0x0022 124 109 000 Old_age Always - 26
196 Reallocated_Event_Count 0x0032 200 200 000 Old_age Always - 0

Samsung 1

188 Command_Timeout 0x0032 100 100 000 Old_age Always - 0
190 Airflow_Temperature_Cel 0x0022 081 078 000 Old_age Always - 19 (Lifetime Min/Max 17/22)
194 Temperature_Celsius 0x0022 081 077 000 Old_age Always - 19 (Lifetime Min/Max 17/23)


Samsung 2

190 Airflow_Temperature_Cel 0x0022 078 076 000 Old_age Always - 22 (Lifetime Min/Max 17/24)
194 Temperature_Celsius 0x0022 078 076 000 Old_age Always - 22 (Lifetime Min/Max 17/24)