BTW, the psio10 dtrace script illuminated the path. When you run a couple of big tar zcvf and everything start to slow down, and psio10 shows "zpool-xxx" 100%, then the system is working only on your snapshots... ;) Gabriele ---------------------------------------------------------------------------------------- Sonicle S.r.l. : http://www.sonicle.com Music: http://www.gabrielebulfon.com Quantum Mechanics : http://www.cdbaby.com/cd/gabrielebulfon Da: Gabriele Bulfon A: disc...@lists.illumos.org Data: 27 aprile 2016 0.57.27 CEST Oggetto: Solved: [discuss] Urgent help on raid zpool on adaptec controller After a lot of dtracing etc I could find the reason for all the slow down on that zpool. Never hold too old snapshots, if you want a fast zpool...: two snapshots holding 1 month, and 2 months before were killing the pool. Destroyed, the system is fast as a shark. :) Gabriele ---------------------------------------------------------------------------------------- Sonicle S.r.l. : http://www.sonicle.com Music: http://www.gabrielebulfon.com Quantum Mechanics : http://www.cdbaby.com/cd/gabrielebulfon Da: Gabriele Bulfon A: disc...@lists.illumos.org Data: 26 aprile 2016 16.44.22 CEST Oggetto: [discuss] Urgent help on raid zpool on adaptec controller Hi, I'm having big issues on a cloud server running many zones and deamons. The ones having problem are only the ones on the sas raidz pool running on an adaptec controller: from time to time during the day, all the zones start to slow down, sometimes up to a state of no return, hard to kill processes, hard to shutdown zones, hard to reset the entire machine. At last, I discovered that zones running on rpool (the sata mirror) always respond fast, so now I know the problem is on the raidz pool. What happened, some weeks ago I had to replace one of the raidz disks, and once replaced I discoverd that the disks ar not in jbod mode, but as single adaptec volumes. So I decided this was an opportunity to change the radiz to jbod, starting to set the new disk in jbod mode, and then replaced the faulted disk with the new jbod on the same channel. It worked. I wanted to do the same on the other disk, week by week, transforming the entire pool into a jbod pool, but I had no time. Now, after some weeks, the problem came out as stated above, so my questions: 1- is it bad to have a radiz with 4 adaptec volumes and one jbod? 2- if one of the disks is faulty, how can I find wich one and detach it? Only a long scrub? Will it impact the running machine? 3- if the controller has problems how can I detect it? 4- if I want to replace the controller, can I move the disks to an LSI even if they were in volume mode mixed with jbod? will the zpool be imported? Thanks for any idea.... Gabriele ---------------------------------------------------------------------------------------- Sonicle S.r.l. : http://www.sonicle.com Music: http://www.gabrielebulfon.com Quantum Mechanics : http://www.cdbaby.com/cd/gabrielebulfon illumos-discuss | Archives | Modify Your Subscription
------------------------------------------- illumos-discuss Archives: https://www.listbox.com/member/archive/182180/=now RSS Feed: https://www.listbox.com/member/archive/rss/182180/21175430-2e6923be Modify Your Subscription: https://www.listbox.com/member/?member_id=21175430&id_secret=21175430-6a77cda4 Powered by Listbox: http://www.listbox.com
