June 24, 201412 yr Author I started it over. It had ran 50+ hours and was completed but I couldn't find the logs on /boot. The subdirectory was not there. I did find some out in /var/adm but I think since I restarted it, it wrote over the old ones. Such as life. Anyway, I am wondering if it is normal for the speed to slow down as it gets into the 90% range of completion? Here is what I am currently getting after it running just the pre-read for 10.5 hours...
June 24, 201412 yr Anyway, I am wondering if it is normal for the speed to slow down as it gets into the 90% range of completion? Yes, you are now reading the inside of the platters...its slower.
June 24, 201412 yr Anyway, I am wondering if it is normal for the speed to slow down as it gets into the 90% range of completion? Yes, you are now reading the outside of the platters...its slower. The outside of the platter is the fastest part AFAIK. Assuming a drive is running at a steady RPM, one revolution at the outside of the disc holds a lot more data than one revolution at the inside of the disc.
June 24, 201412 yr Anyway, I am wondering if it is normal for the speed to slow down as it gets into the 90% range of completion? Yes, you are now reading the outside of the platters...its slower. The outside of the platter is the fastest part AFAIK. Assuming a drive is running at a steady RPM, one revolution at the outside of the disc holds a lot more data than one revolution at the inside of the disc. Yes, thank you. I corrected my post, hoping nobody saw it yet.
June 24, 201412 yr My WD Red 4TB's dropped to about 77 mb/s towards the end but started of at 147 mb/s. Perfectly normal Pre-Read, my temp's on the drives normally max out at 34c during the preclear.
June 24, 201412 yr Sounds just about right for Reds -- I had about the same performance with mine. I did notice a slight difference in timings if they were done on my Atom server vs. an i5-based server -- but it was only a few minutes out of the many hours of pre-clear times, so it could have even been just a variance in drives.
June 25, 201412 yr Author Ok, it is currently on the Post-Read and is 15% complete. Here is what the "smart_mid_after_zer01_sde" report looks like. Can someone interpret this for me and let me know if there is not enough data to go on yet or tell me if this disk is a no-go and I should take it back today? SMART Attributes Data Structure revision number: 10 Vendor Specific SMART Attributes with Thresholds: ID# ATTRIBUTE_NAME FLAG VALUE WORST THRESH TYPE UPDATED WHEN_FAILED RAW_VALUE 1 Raw_Read_Error_Rate 0x000f 119 100 006 Pre-fail Always - 222532928 3 Spin_Up_Time 0x0003 092 091 000 Pre-fail Always - 0 4 Start_Stop_Count 0x0032 100 100 020 Old_age Always - 21 5 Reallocated_Sector_Ct 0x0033 100 100 010 Pre-fail Always - 0 7 Seek_Error_Rate 0x000f 075 057 030 Pre-fail Always - 8664241076 9 Power_On_Hours 0x0032 100 100 000 Old_age Always - 331 10 Spin_Retry_Count 0x0013 100 100 097 Pre-fail Always - 0 12 Power_Cycle_Count 0x0032 100 100 020 Old_age Always - 18 183 Runtime_Bad_Block 0x0032 099 099 000 Old_age Always - 1 184 End-to-End_Error 0x0032 100 100 099 Old_age Always - 0 187 Reported_Uncorrect 0x0032 100 100 000 Old_age Always - 0 188 Command_Timeout 0x0032 100 099 000 Old_age Always - 4295032834 189 High_Fly_Writes 0x003a 097 097 000 Old_age Always - 3 190 Airflow_Temperature_Cel 0x0022 065 063 045 Old_age Always - 35 (Min/Max 28/36) 194 Temperature_Celsius 0x0022 035 040 000 Old_age Always - 35 (0 25 0 0 0) 195 Hardware_ECC_Recovered 0x001a 119 100 000 Old_age Always - 222532928 197 Current_Pending_Sector 0x0012 100 100 000 Old_age Always - 0 198 Offline_Uncorrectable 0x0010 100 100 000 Old_age Offline - 0 199 UDMA_CRC_Error_Count 0x003e 200 200 000 Old_age Always - 0 240 Head_Flying_Hours 0x0000 100 253 000 Old_age Offline - 271111220625721 241 Total_LBAs_Written 0x0000 100 253 000 Old_age Offline - 33607716198 242 Total_LBAs_Read 0x0000 100 253 000 Old_age Offline - 52063306179 SMART Error Log Version: 1 No Errors Logged
June 25, 201412 yr Author The Raw_Read_Error_Rate, Seek_Error_Rate, and Command_Timeout seem awfully high.
June 25, 201412 yr The Raw_Read_Error_Rate, Seek_Error_Rate, and Command_Timeout seem awfully high. Well...it is a Seagate!
June 25, 201412 yr The Raw_Read_Error_Rate, Seek_Error_Rate, and Command_Timeout seem awfully high. With a few exceptions the RAW attribute values are not useful to us. The value is 32 bits or something and each manufacturer can use those bits for anything they want. So the first 5 bits may mean one thing, the next 3 something else, etc. When you string it all together and make a number of it it means nothing. You really need to look at the standardized values and see if they are approaching the threshold. The ones that we do monitor the RAW values are reallocated sectors, pending sectors, UDMA/ECC errors and temperature. Other attributes like power on hours are also readable and accurate (at least for all drives I.have seen). I see nothing concerning about the smart report you posted. Seagate drives have not been the highest.quality in recent years IMO and several studies bear that out. But the recent generation of drives have faired better and I.have bought several 4T and 5T drives.and (fingers crossed) so far so good. You are right that drives from any manufacturers can and do fail.
June 25, 201412 yr Meaning what? I have had WD green 1TB and Hitachi 1TB fail too. I've had over 20 Seagate failures in the last few years. Not a single HGST/Hitachi or WD failure. They are cheaper for a reason. See BackBlaze's stats: http://blog.backblaze.com/2014/01/21/what-hard-drive-should-i-buy/
June 25, 201412 yr Author Meaning what? I have had WD green 1TB and Hitachi 1TB fail too. I've had over 20 Seagate failures in the last few years. Not a single HGST/Hitachi or WD failure. They are cheaper for a reason. See BackBlaze's stats: http://blog.backblaze.com/2014/01/21/what-hard-drive-should-i-buy/ I do like stats as I work for a storage company. If anyone would know, it would be Backblaze. With this information, my next ones will be Hitachi 4TB disks. I have a bunch of 1TB ones to replace, even though they have not failed.
June 25, 201412 yr Oh god, not this blackblaze study being tossed around again. It has been discussed multiple times in multiple threads. It is not the end all be all once you realize the limitations that study has.
June 25, 201412 yr Oh god, not this blackblaze study being tossed around again. It has been discussed multiple times in multiple threads. It is not the end all be all once you realize the limitations that study has. It was merely an example. There have been LOTS of Seagates going bad the last few years.
June 25, 201412 yr Right, but which model,which capacity? To blindly group them all together at the manufacturer level is like saying "all cars are bad because of recalls" when in reality it is a specific make and model and production factory. Details do matter.
June 26, 201412 yr Author Brit is right here. I work for a company that sells data storage. Hitachi and Seagate are our two used vendors. I can tell you that disks with the SMOOS in their model from Seagate, run far away from. They are known to have huge problems. They are called "Moose" drives.
June 26, 201412 yr Brit is right here. I work for a company that sells data storage. Hitachi and Seagate are our two used vendors. I can tell you that disks with the SMOOS in their model from Seagate, run far away from. They are known to have huge problems. They are called "Moose" drives. I think the drives with model ending with DM000 are the ones that are getting more favorable responses from BackBlaze (not BlackBlaze ).
June 26, 201412 yr Author Yeah, I think those are Moose drives too. Apple even had me bring in a iMac I had last year because they needed to replace the hdd in it. I knew what was up.
June 26, 201412 yr Author Yeah, I think those are Moose drives too. Apple even had me bring in a iMac I had last year because they needed to replace the hdd in it. I knew what was up. I stand corrected. I see you said those were "favorable" by backblaze. So yes, those are probably not Moose drives.
June 26, 201412 yr Right, but which model,which capacity? To blindly group them all together at the manufacturer level is like saying "all cars are bad because of recalls" when in reality it is a specific make and model and production factory. Details do matter. The Backblaze article that you dismissed broke it down by make and specific model. Seagates came out the worst overall, but clearly from Backblaze's study, there were some winners and losers in each individual lineup. Also, can you point to another article that actually names manufacturers and specific models *and* that has a statistical significance like Backblaze's study of 25K drives? It may not be a perfect study, in that the data can't be applied elsewhere exactly as in their data center and usage scenario, but it's the only mass-scale study that I know of that "named names." Google's white paper from 2007 wussed out and didn't name specific drive models nor manufacturers, despite stating that there were important correlations. Backblaze seems to be the only ones who have the gumption to do that. The Backblaze study may not be perfect - but it's better than nearly everything else out there right now.
June 26, 201412 yr Author Right, but which model,which capacity? To blindly group them all together at the manufacturer level is like saying "all cars are bad because of recalls" when in reality it is a specific make and model and production factory. Details do matter. The Backblaze article that you dismissed broke it down by make and specific model. Seagates came out the worst overall, but clearly from Backblaze's study, there were some winners and losers in each individual lineup. Also, can you point to another article that actually names manufacturers and specific models *and* that has a statistical significance like Backblaze's study of 25K drives? It may not be a perfect study, in that the data can't be applied elsewhere exactly as in their data center and usage scenario, but it's the only mass-scale study that I know of that "named names." Google's white paper from 2007 wussed out and didn't name specific drive models nor manufacturers, despite stating that there were important correlations. Backblaze seems to be the only ones who have the gumption to do that. The Backblaze study may not be perfect - but it's better than nearly everything else out there right now. I agree with you, InterFaceLift.
Archived
This topic is now archived and is closed to further replies.