From: Frieder F. <fri...@we...> - 2010-03-15 22:43:40
|
Hi, just had a look into the output of smartd in /var/log/messages. Seems like the log mistakenly contains 198 Read_Sectors_Tot_Ct as Offline uncorrectable sectors? Mar 15 17:43:51 lake smartd[2821]: smartd 5.40 2010-03-05 r3074 [x86_64-suse-linux-gnu] (local build)#012Copyright (C) 2002-10 by Bruce Allen, http://smartmontools.sourceforge.net#012 Mar 15 17:43:51 lake smartd[2821]: No configuration file /usr/local/etc/smartd.conf found, scanning devices ... Mar 15 17:43:51 lake smartd[3116]: smartd has fork()ed into background mode. New PID=3116. ... Mar 15 18:03:50 lake rsyslogd: -- MARK -- Mar 15 18:13:51 lake smartd[3116]: Device: /dev/sda [SAT], 95840173 Offline uncorrectable sectors (changed +703184) ... Mar 15 18:43:50 lake rsyslogd: -- MARK -- Mar 15 18:43:51 lake smartd[3116]: Device: /dev/sda [SAT], 96170325 Offline uncorrectable sectors (changed +330152) ... Mar 15 19:13:51 lake smartd[3116]: Device: /dev/sda [SAT], 96449717 Offline uncorrectable sectors (changed +279392) ... Mar 15 19:43:51 lake smartd[3116]: Device: /dev/sda [SAT], 96457509 Offline uncorrectable sectors (changed +7792) ... It would be really nice if: 208 Average_Erase_Count 209 Remaining_Lifetime_Perc would instead be reported in the log:) Am 09.03.2010 21:09, schrieb Frieder Ferlemann: > actually I did never expect to report back this early, > but I'm quite upset about the Average Erase Count > of 72 of the drive. > > 198 Read_Sectors_Tot_Ct 0x0000 --- --- --- Old_age Offline - 82208608 > 199 Write_Sectors_Tot_Ct 0x0000 --- --- --- Old_age Offline - 106676155 > 200 Read_Commands_Tot_Ct 0x0000 --- --- --- Old_age Offline - 1781513 > 201 Write_Commands_Tot_Ct 0x0000 --- --- --- Old_age Offline - 2275882 > > 208 Average_Erase_Count 0x0000 --- --- --- Old_age Offline - 72 > > Meanwhile 6492 MByte have been written (106676155 - 93996899) Sectors * 512 Bytes /Sector > while 1408511 MByte have been erased! (72 - 50) * 64.023.257.088 Bytes > > This would mean that wear leaveling (or garbage collection/whatever) > causes an overhead of more than 200 in flash erase cycles!!! > > Intolerable. > > I hope there is a mistake in my calculations or somewhere else. meanwhile the drive seems to continue self destruction: 9 Power_On_Hours 0x0000 --- --- --- Old_age Offline - 126 12 Power_Cycle_Count 0x0000 --- --- --- Old_age Offline - 36 198 Read_Sectors_Tot_Ct 0x0000 --- --- --- Old_age Offline - 96638645 199 Write_Sectors_Tot_Ct 0x0000 --- --- --- Old_age Offline - 115111795 200 Read_Commands_Tot_Ct 0x0000 --- --- --- Old_age Offline - 2073793 201 Write_Commands_Tot_Ct 0x0000 --- --- --- Old_age Offline - 2531426 208 Average_Erase_Count 0x0000 --- --- --- Old_age Offline - 88 209 Remaining_Lifetime_Perc 0x0000 --- --- --- Old_age Offline - 100 Since the last post 4319 MByte have been written (115111795 - 106676155) Sectors * 512 Bytes /Sector while 1024372 MByte have been erased! (88 - 72) * 64.023.257.088 Bytes So again more than 200 times more have been erased than have been written!!! (output of smartctl -a is appended). To put this factor into perspective: For some other SSD drives "write amplification multipliers" of as low as 1.1 have been reported/claimed! (I'd be happy with a factor of say 4 to 8 ) And I'm trying to treat the drive like a raw egg: - More than 5% of the total drive capacity were never assigned to a partition. - And were never written to. - And the root partition was mounted noatime (output of "mount" was shown in an earlier post). - And a 2.6.33 kernel is used. - And I have aligned the partitions to 2 MBytes: > fdisk -l /dev/sda Disk /dev/sda: 64.0 GB, 64023257088 bytes 128 heads, 32 sectors/track, 30528 cylinders Units = cylinders of 4096 * 512 = 2097152 bytes Disk identifier: 0x3d2797a4 Device Boot Start End Blocks Id System /dev/sda1 2 16386 33556480 83 Linux /dev/sda2 * 16387 28675 25167872 83 Linux Greetings, Frieder |