Badblocks
in [Linux Hardware]
|
From: General Schvantzkoph on 14 Jul 2010 12:38 I have some questions about badblocks. I'm running badblocks on a new WD 1.5TB drive. On Storage Review the performance of the Caviar Black series for sequential reads and writes is measured at 75MBytes/sec to 145MBytes/ sec. SATA is half duplex so the expected time to read and write all 1.5TBytes would be around 8 hours (assuming an average of 110MB/second for each operate, i.e. 55MBytes/second for Write+Read). However the progress indicator from badblocks is indicating that it will take about 3 hours for one pass which is much faster than expected. Does anyone know if badblocks is testing all of the blocks or is it just testing a subset? I've told badblocks to test 1GB at a time (the system has 8G of RAM), is it possible that the drive is really achieving a write+read speed of 140MBytes/second? badblocks -b 16384 -c 65536 -v -s -w /dev/sdc Checking for bad blocks in read-write mode From block 0 to 91571160 Testing with pattern 0xaa: 25.62% done, 45:37 elapsed
From: General Schvantzkoph on 14 Jul 2010 15:46 On Wed, 14 Jul 2010 16:38:48 +0000, General Schvantzkoph wrote: > I have some questions about badblocks. I'm running badblocks on a new WD > 1.5TB drive. On Storage Review the performance of the Caviar Black > series for sequential reads and writes is measured at 75MBytes/sec to > 145MBytes/ sec. SATA is half duplex so the expected time to read and > write all 1.5TBytes would be around 8 hours (assuming an average of > 110MB/second for each operate, i.e. 55MBytes/second for Write+Read). > However the progress indicator from badblocks is indicating that it will > take about 3 hours for one pass which is much faster than expected. > > Does anyone know if badblocks is testing all of the blocks or is it just > testing a subset? I've told badblocks to test 1GB at a time (the system > has 8G of RAM), is it possible that the drive is really achieving a > write+read speed of 140MBytes/second? > > badblocks -b 16384 -c 65536 -v -s -w /dev/sdc Checking for bad blocks in > read-write mode From block 0 to 91571160 > Testing with pattern 0xaa: 25.62% done, 45:37 elapsed Never mind. badblocks treats the writes and reads as separate phases, the time estimate was for the write phase only followed by a read phase. Those times make sense. badblocks -b 16384 -c 65536 -v -s -w /dev/sdc Checking for bad blocks in read-write mode From block 0 to 91571160 Testing with pattern 0xaa: done Reading and comparing: 14.24% done, 4:01:18 elapsed
From: Stefan Patric on 15 Jul 2010 12:16 On Wed, 14 Jul 2010 19:46:36 +0000, General Schvantzkoph wrote: > On Wed, 14 Jul 2010 16:38:48 +0000, General Schvantzkoph wrote: > >> I have some questions about badblocks. I'm running badblocks on a new >> WD 1.5TB drive. On Storage Review the performance of the Caviar Black >> series for sequential reads and writes is measured at 75MBytes/sec to >> 145MBytes/ sec. SATA is half duplex so the expected time to read and >> write all 1.5TBytes would be around 8 hours (assuming an average of >> 110MB/second for each operate, i.e. 55MBytes/second for Write+Read). >> However the progress indicator from badblocks is indicating that it >> will take about 3 hours for one pass which is much faster than >> expected. >> >> Does anyone know if badblocks is testing all of the blocks or is it >> just testing a subset? I've told badblocks to test 1GB at a time (the >> system has 8G of RAM), is it possible that the drive is really >> achieving a write+read speed of 140MBytes/second? >> >> badblocks -b 16384 -c 65536 -v -s -w /dev/sdc Checking for bad blocks >> in read-write mode From block 0 to 91571160 Testing with pattern 0xaa: >> 25.62% done, 45:37 elapsed > > Never mind. badblocks treats the writes and reads as separate phases, > the time estimate was for the write phase only followed by a read phase. > Those times make sense. > > badblocks -b 16384 -c 65536 -v -s -w /dev/sdc Checking for bad blocks in > read-write mode From block 0 to 91571160 > Testing with pattern 0xaa: done > Reading and comparing: 14.24% done, 4:01:18 elapsed If it's any help, I just finished doing a badblock check (defaults and -w option) of an older 7200RPM 40GB IDE drive. It took 3 hours and 20 minutes to complete the check. Stef
From: General Schvantzkoph on 15 Jul 2010 17:16 On Thu, 15 Jul 2010 16:16:06 +0000, Stefan Patric wrote: > On Wed, 14 Jul 2010 19:46:36 +0000, General Schvantzkoph wrote: > >> On Wed, 14 Jul 2010 16:38:48 +0000, General Schvantzkoph wrote: >> >>> I have some questions about badblocks. I'm running badblocks on a new >>> WD 1.5TB drive. On Storage Review the performance of the Caviar Black >>> series for sequential reads and writes is measured at 75MBytes/sec to >>> 145MBytes/ sec. SATA is half duplex so the expected time to read and >>> write all 1.5TBytes would be around 8 hours (assuming an average of >>> 110MB/second for each operate, i.e. 55MBytes/second for Write+Read). >>> However the progress indicator from badblocks is indicating that it >>> will take about 3 hours for one pass which is much faster than >>> expected. >>> >>> Does anyone know if badblocks is testing all of the blocks or is it >>> just testing a subset? I've told badblocks to test 1GB at a time (the >>> system has 8G of RAM), is it possible that the drive is really >>> achieving a write+read speed of 140MBytes/second? >>> >>> badblocks -b 16384 -c 65536 -v -s -w /dev/sdc Checking for bad blocks >>> in read-write mode From block 0 to 91571160 Testing with pattern 0xaa: >>> 25.62% done, 45:37 elapsed >> >> Never mind. badblocks treats the writes and reads as separate phases, >> the time estimate was for the write phase only followed by a read >> phase. Those times make sense. >> >> badblocks -b 16384 -c 65536 -v -s -w /dev/sdc Checking for bad blocks >> in read-write mode From block 0 to 91571160 Testing with pattern 0xaa: >> done >> Reading and comparing: 14.24% done, 4:01:18 elapsed > > If it's any help, I just finished doing a badblock check (defaults and > -w option) of an older 7200RPM 40GB IDE drive. It took 3 hours and 20 > minutes to complete the check. > > Stef I figure badblocks will be done in another hour or two, approximately 32 hours total to run in on a 1.5T drive.
From: HASM on 15 Jul 2010 19:37 General Schvantzkoph <schvantzkoph(a)yahoo.com> writes: > I figure badblocks will be done in another hour or two, approximately 32 > hours total to run in on a 1.5T drive. Yup, these drives are huge. It takes just as long to randomize it for a luks encryption partition ... -- HASM
|
Pages: 1 Prev: Low level formatting of laser servo diskettes Next: Cirrus Alpine Video RAM |