Bcache SSD life



  • Does anyone have any figures for bcache? I am trying to decide which SSD to go with but need to get an idea of how much and how often bcache writes to the cache drive. If it's low enough I could get away with buying a cheaper consumer drive, otherwise I'm looking into getting a Samsung 840 Pro.



  • If the information is of use to anyone: I previously used EnhanceIO but saw little to no improvement in random IO (over NFS). The cache drive quit after about 3 months, it was one of those cheap PNY ones run through an IBM M1015. Oddly enough the cache never seemed to write more than a few gigabytes to the drive, even with writeback enabled. The drive became visible again after having been left off for a few months, I haven't experimented with it to see how healthy it is yet.



  • Tried giving the drive a second chance, it lasted a couple of days before dying again. Turns out the PNY XLR8s have Sandforce controllers, I wouldn't be surprised if PNY didn't update the firmware before shipping them out.



  • That is crazy that they are dying so quickly!



  • @rd146352 said:

    If the information is of use to anyone: I previously used EnhanceIO but saw little to no improvement in random IO (over NFS). The cache drive quit after about 3 months, it was one of those cheap PNY ones run through an IBM M1015. Oddly enough the cache never seemed to write more than a few gigabytes to the drive, even with writeback enabled. The drive became visible again after having been left off for a few months, I haven't experimented with it to see how healthy it is yet.

    I've killed cheap ADATA ones doing similar. They may be fine for the average punter but you put some real stress on them and they go squish. Get the Samsung's.



  • Ended up revising my approach to the whole configuration. I bought an SM951 128GB for the caching device and grabbed a pair of seagate 2TB drives and put them in an md raid 10. Random 4k inside one of the VMs averages between 12,000 and 20,000 write depending on queue depth. This seems low given that the SM951 is rated really high, can bcache be tuned to increase the dirty cache for random io?

    I did manage to find a setting for sequential which pushed it up to 1GB/s.