We are running FreeBSD 10.1 and we had 4 1 TB SSDs in L2ARC which we removed and made into a separate zpool. I ran sysctl and noticed a strange message showing this lost memory. I have pasted output of sysctl -a |grep UMA. Any idea why this is happening ? Any fixes or suggested solution ? hosts (1/10): sjc1-prod1-01.voxer.com Freed UMA keg (zio_data_buf_1024) was not empty (333 items). Lost 171 pages of memory. Freed UMA keg (zio_data_buf_1536) was not empty (138 items). Lost 583 pages of memory. Freed UMA keg (zio_data_buf_2048) was not empty (84 items). Lost 148 pages of memory. Freed UMA keg (zio_data_buf_1024) was not empty (380 items). Lost 175 pages of memory. Freed UMA keg (zio_data_buf_1536) was not empty (154 items). Lost 328 pages of memory. Freed UMA keg (zio_data_buf_2048) was not empty (67 items). Lost 91 pages of memory. Freed UMA keg (zio_data_buf_1024) was not empty (256 items). Lost 95 pages of memory. Freed UMA keg (zio_data_buf_1536) was not empty (170 items). Lost 383 pages of memory. Freed UMA keg (zio_data_buf_2048) was not empty (117 items). Lost 257 pages of memory. hosts (1/10): sjc1-prod1-02.voxer.com Freed UMA keg (zio_data_buf_1024) was not empty (336 items). Lost 175 pages of memory. Freed UMA keg (zio_data_buf_1536) was not empty (157 items). Lost 638 pages of memory. Freed UMA keg (zio_data_buf_2048) was not empty (83 items). Lost 150 pages of memory. Freed UMA keg (zio_data_buf_1024) was not empty (329 items). Lost 140 pages of memory. Freed UMA keg (zio_data_buf_1536) was not empty (216 items). Lost 425 pages of memory. Freed UMA keg (zio_data_buf_2048) was not empty (85 items). Lost 130 pages of memory. Freed UMA keg (zio_data_buf_1024) was not empty (437 items). Lost 173 pages of memory. Freed UMA keg (zio_data_buf_1536) was not empty (165 items). Lost 576 pages of memory. Freed UMA keg (zio_data_buf_2048) was not empty (103 items). Lost 273 pages of memory. Thanks, George
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA512 On 12/04/14 16:23, George Kola wrote:> We are running FreeBSD 10.1 and we had 4 1 TB SSDs in L2ARC which > we removed and made into a separate zpool. I ran sysctl and noticed > a strange message showing this lost memory. I have pasted output of > sysctl -a |grep UMA. Any idea why this is happening ? Any fixes or > suggested solution ?This may be fixed by r275492 of stable/10. There is a known memory leak in L2ARC code when compression is used. Cheers, - -- Xin LI <delphij at delphij.net> https://www.delphij.net/ FreeBSD - The Power to Serve! Live free or die -----BEGIN PGP SIGNATURE----- Version: GnuPG v2.1.0 (FreeBSD) iQIcBAEBCgAGBQJUgP19AAoJEJW2GBstM+nsLm4P/RlU/MaNdaUjO447/Anw/5Ok 3GKjVfFP8RbOyO7LyG6lMSOhXS2OWLqFXwLBthF3ISdOKT7pbJLw0Dus6FEJAAYh tU7eIWBsLfiEaloxQmC8J3AutFLRFYkY01Vq3Vh4zeLJoYfbOvSA8Z7d/KYgDOiw Kkjs8uROf7+Op3zx26C4mdryw8D8yxyq6kmbuJwQ6VofblqFu2UtGt5lpXuuJLLh OfixO+mg61hy8hNRLsEZCirSOAWatplha0rdu15Gi/wr/mCSWvi22u/IZ7mmp5di aIouGy/1g6VXoINFpGEHmFX395/NE4bPep3sDhy/TEI8S2hPxJf0bXy99SzRTbiy YPMH1rxougcQXjP27gu2nC6IAWNFQiTJo6h6cV+2XReeMC21IeEaz+yrvWV64AJ1 iNGHTS+RKtgrTjNdYsGdX1OQ+H+s0p4BUf/8MSyhakSJzCJS8wnhet4Qj2+YpfwX SJtG5Ox7TKTJSfmMRJh95aP1uS3I75xNmCmbLrmQ0ytFfjoD4jjz908gc/zwyb3t /pWlzdqLS1Paw8cn+yryrc4uMd0MRGnyOHMKL1Ut00VFjtUrFXhuqkdjDBBBtzzM kHINlkCQ7cfCBnA6T6RvobGZ2MdGNgXnAiahLOtXqTBK3W4Btfqd5JX8WfVlXtwL ONzglteBql7rQZWpEm/x =qzsA -----END PGP SIGNATURE-----