smartd, mfi, SAS and SATA
Hi everyone, I recently took delivery of a Supermicro X11SRM-F with a Broadcom MegaRAID 9361-8i SAS 8 port card which has 4 Intel D3-S4610 960 GB SSDs and 4 Hitachi/WD Ultrastar HC300 4TB drives each in a RAID5. I have /usr/local/etc/smartd.conf with just 'DEVICESCAN' and when smartd starts I see.. Nov 21 01:17:49 maarsy-acq3 smartd[2103]: Opened configuration file /usr/local/etc/smartd.conf Nov 21 01:17:49 maarsy-acq3 smartd[2103]: Drive: DEVICESCAN, implied '-a' Directive on line 23 of file /usr/local/etc/smartd.conf Nov 21 01:17:49 maarsy-acq3 smartd[2103]: Configuration file /usr/local/etc/smartd.conf was parsed, found DEVICESCAN, scanning devices Nov 21 01:17:49 maarsy-acq3 smartd[2103]: Device: /dev/pass0, opened Nov 21 01:17:49 maarsy-acq3 smartd[2103]: Device: /dev/pass0, [HGST HUS726T4TAL5204 C40H], lu id: 0x5000cca097502308, S/N: V6HE27MR, 4.00 TB Nov 21 01:17:50 maarsy-acq3 smartd[2103]: Device: /dev/pass0, is SMART capable. Adding to "monitor" list. Nov 21 01:17:50 maarsy-acq3 smartd[2103]: Device: /dev/pass1, opened Nov 21 01:17:50 maarsy-acq3 smartd[2103]: Device: /dev/pass1, [HGST HUS726T4TAL5204 C40H], lu id: 0x5000cca09751d2c8, S/N: V6HEZZZR, 4.00 TB Nov 21 01:17:50 maarsy-acq3 smartd[2103]: Device: /dev/pass1, is SMART capable. Adding to "monitor" list. Nov 21 01:17:50 maarsy-acq3 smartd[2103]: Device: /dev/pass2, opened Nov 21 01:17:50 maarsy-acq3 smartd[2103]: Device: /dev/pass2, [HGST HUS726T4TAL5204 C40H], lu id: 0x5000cca097509ad8, S/N: V6HEA6ZR, 4.00 TB Nov 21 01:17:51 maarsy-acq3 smartd[2103]: Device: /dev/pass2, is SMART capable. Adding to "monitor" list. Nov 21 01:17:51 maarsy-acq3 smartd[2103]: Device: /dev/pass3, opened Nov 21 01:17:51 maarsy-acq3 smartd[2103]: Device: /dev/pass3, [HGST HUS726T4TAL5204 C40H], lu id: 0x5000cca0974f1630, S/N: V6HDHALR, 4.00 TB Nov 21 01:17:51 maarsy-acq3 smartd[2103]: Device: /dev/pass3, is SMART capable. Adding to "monitor" list. Nov 21 01:17:51 maarsy-acq3 smartd[2103]: Device: /dev/pass4, type changed from 'scsi' to 'sat' Nov 21 01:17:51 maarsy-acq3 smartd[2103]: Device: /dev/pass5, type changed from 'scsi' to 'sat' Nov 21 01:17:51 maarsy-acq3 smartd[2103]: Device: /dev/pass6, type changed from 'scsi' to 'sat' Nov 21 01:17:52 maarsy-acq3 smartd[2103]: Device: /dev/pass7, type changed from 'scsi' to 'sat' Nov 21 01:17:52 maarsy-acq3 smartd[2103]: Monitoring 0 ATA/SATA, 4 SCSI/SAS and 0 NVMe devices Nov 21 01:17:53 maarsy-acq3 smartd[2105]: smartd has fork()ed into background mode. New PID=2105. Nov 21 01:17:53 maarsy-acq3 smartd[2105]: file /var/run/smartd.pid written containing PID 2105 So it is monitoring the SAS disks but has ignored the SATA SSDs :( [maarsy-acq3 1:33] ~> camcontrol devlist at scbus8 target 8 lun 0 (pass0) at scbus8 target 9 lun 0 (pass1) at scbus8 target 10 lun 0 (pass2) at scbus8 target 11 lun 0 (pass3) at scbus8 target 12 lun 0 (pass4) at scbus8 target 13 lun 0 (pass5) at scbus8 target 14 lun 0 (pass6) at scbus8 target 15 lun 0 (pass7) at scbus9 target 0 lun 0 (da0,pass8) If I run smartctl on an SSD I get.. [maarsy-acq3 1:33] ~> sudo smartctl -a /dev/pass4|less smartctl 7.0 2018-12-30 r4883 [FreeBSD 12.0-RELEASE amd64] (local build) Copyright (C) 2002-18, Bruce Allen, Christian Franke, www.smartmontools.org Smartctl open device: /dev/pass4 [SAT] failed: SATA device detected, MegaRAID SAT layer is reportedly buggy, use '-d sat' to try anyhow And using what it suggests seems to work - does anyone know a way to force it to work with DEVICESCAN? For now I've just hard coded it like so.. DEFAULT -m root /dev/pass0 /dev/pass1 /dev/pass2 /dev/pass3 /dev/pass4 -d sat /dev/pass5 -d sat /dev/pass6 -d sat /dev/pass7 -d sat but it seems clunky.. also I see these slightly puzzling messages for each SSD.. Nov 21 01:37:08 maarsy-acq3 smartd[3656]: Device: /dev/pass4 [SAT], opened Nov 21 01:37:08 maarsy-acq3 smartd[3656]: Device: /dev/pass4 [SAT], INTEL SSDSC2KB960G8, S/N:PHYF92630636960CGN, WWN:5-5cd2e4-150f430c3, FW:XCV10120, 960 GB Nov 21 01:37:08 maarsy-acq3 smartd[3656]: Device: /dev/pass4 [SAT], found in smartd database: Intel S4510/S4610/S4500/S4600 Series SSDs Nov 21 01:37:08 maarsy-acq3 smartd[3656]: Device: /dev/pass4 [SAT], not capable of SMART Health Status check Nov 21 01:37:08 maarsy-acq3 smartd[3656]: Device: /dev/pass4 [SAT], can't monitor Offline_Uncorrectable count - no Attribute 198 Nov 21 01:37:08 maarsy-acq3 smartd[3656]: Device: /dev/pass4 [SAT], is SMART capable. Adding to "monitor" list. Which I am hoping aren't anything to worry about.. -- Daniel O'Connor "The nice thing about standards is that there are so many of them to choose from." -- Andrew Tanenbaum ___ freebsd-stable@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/freebsd-stable To unsubscribe, send any mail to "freebsd-stable-unsubscr...@freebsd.org"
12.0->12.1 and beadm/bectl issues
Hi, After upgrading from 12.0-RELEASE-p11 to 12.1-RELEASE I was having some issues with kld_load and linux support which, after searching [1], seemed due to a missing /boot folder after the upgrade. This was fixed with 'ln -s /bootpool/boot /boot'. Then yesterday when I was trying to switch from quarterly packages to latest I wanted to use a new boot environment and so went through the beadm create and beadm activate but it wouldn't activate with a zpool.cache cp message and it left the new BE mounted under /tmp. After umounting and destroying I repeated the process with bectl and it worked fine, however, upon reboot I was not in the new BE but the same BE and the new one was still marked as activated for use next boot. So firstly: are the be* issues related to the earlier upgrade fix? Secondly: shouldn't beadm and bectl behave the same? Thirdly: how can I properly activate and boot to a new BE? Below is the command output of the beadm/bectl process described above. If there's any more information I can provide please let me know. Thanks, Jon [1] https://forums.freebsd.org/threads/cannot-identify-running-kernel-after-upgrading-to-freebsd-12.68772/ This is a two disk mirrored zpool on GELI with encrypted swap as configured out of the box by the 12.0 installer. root@prometheus:~ # uname -a FreeBSD prometheus 12.1-RELEASE-p1 FreeBSD 12.1-RELEASE-p1 GENERIC amd64 root@prometheus:~ # beadm list BEActive Mountpoint Space Created 12_0-RELEASE-p11 - -1.1G 2019-10-29 21:33 12_1-RELEASE-p1-quarterly NR / 32.7G 2019-11-05 22:24 root@prometheus:~ # beadm create test Created successfully root@prometheus:~ # beadm list BEActive Mountpoint Space Created 12_0-RELEASE-p11 - -1.1G 2019-10-29 21:33 12_1-RELEASE-p1-quarterly NR / 32.7G 2019-11-05 22:24 test - -8.0K 2019-11-20 13:24 root@prometheus:~ # beadm activate test cp: /tmp/BE-test.pJtR9Rs6/boot/zfs/zpool.cache and /boot/zfs/zpool.cache are identical (not copied). root@prometheus:~ # beadm list BEActive Mountpoint Space Created 12_0-RELEASE-p11 - - 1.1G 2019-10-29 21:33 12_1-RELEASE-p1-quarterly NR / 32.7G 2019-11-05 22:24 test - /tmp/BE-test.pJtR9Rs6 136.0K 2019-11-20 13:24 root@prometheus:~ # beadm umount test Unmounted successfully root@prometheus:~ # beadm list BEActive Mountpoint Space Created 12_0-RELEASE-p11 - -1.1G 2019-10-29 21:33 12_1-RELEASE-p1-quarterly NR / 32.7G 2019-11-05 22:24 test - - 136.0K 2019-11-20 13:24 root@prometheus:~ # beadm destroy test Are you sure you want to destroy 'test'? This action cannot be undone (y/[n]): y Destroyed successfully root@prometheus:~ # beadm list BEActive Mountpoint Space Created 12_0-RELEASE-p11 - -1.1G 2019-10-29 21:33 12_1-RELEASE-p1-quarterly NR / 32.7G 2019-11-05 22:24 root@prometheus:~ # bectl list BEActive Mountpoint Space Created 12_0-RELEASE-p11 - - 1.14G 2019-10-29 21:33 12_1-RELEASE-p1-quarterly NR / 32.7G 2019-11-05 22:24 root@prometheus:~ # bectl create test root@prometheus:~ # bectl list BEActive Mountpoint Space Created 12_0-RELEASE-p11 - - 1.14G 2019-10-29 21:33 12_1-RELEASE-p1-quarterly NR / 32.7G 2019-11-05 22:24 test - - 8K2019-11-20 13:25 root@prometheus:~ # bectl activate test successfully activated boot environment test root@prometheus:~ # bectl list BEActive Mountpoint Space Created 12_0-RELEASE-p11 - - 1.14G 2019-10-29 21:33 12_1-RELEASE-p1-quarterly N / 8K2019-11-05 22:24 test R - 32.7G 2019-11-20 13:25 root@prometheus:~ # beadm list BEActive Mountpoint Space Created 12_0-RELEASE-p11 - -1.1G 2019-10-29 21:33 12_1-RELEASE-p1-quarterly N /8.0K 2019-11-05 22:24 test R - 32.7G 2019-11-20 13:25 root@prometheus:~ # Following a reboot I'll still be running in 12_1-RELEASE-p1-quarterly and test will still be marked R. ___ freebsd-stable@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/freebsd-stable To unsubscribe, send any mail to "freebsd-stable-unsubscr...@freebsd.org"
Re: jexec as user?
Yeah, ssh is also possible. See my original mail :-) https://lists.freebsd.org/pipermail/freebsd-stable/2019-November/091742.html I run my jails with "ip4 = inherit;". So I would need to do some port-forwarding trickery with ssh on different ports. The users already login on the host to do various actions. Jailme gives the easiest access without to much maintenance for now. Regards, Ronald. Van: Eugene Grosbein Datum: woensdag, 20 november 2019 11:44 Aan: Ronald Klop , Miroslav Lachman <000.f...@quip.cz> CC: Christos Chatzaras , freebsd-stable Onderwerp: Re: jexec as user? 20.11.2019 16:47, Ronald Klop wrote: > Thanks for all the advice. I am indeed looking for using jail from the non-root user in the host. Jailme sounds like a good solution. > > My use case is providing a relatively save way of giving a user the possibility to experiment with root rights (like creating and installing ports) without wracking the host system. > The users are trusted so it is not so much about security. More about keeping the host system clean. You also could run ssh service inside the jail and give users opportunity to experiment with ssh and keys :-) ___ freebsd-stable@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/freebsd-stable To unsubscribe, send any mail to "freebsd-stable-unsubscr...@freebsd.org"
Re: jexec as user?
20.11.2019 16:47, Ronald Klop wrote: > Thanks for all the advice. I am indeed looking for using jail from the > non-root user in the host. Jailme sounds like a good solution. > > My use case is providing a relatively save way of giving a user the > possibility to experiment with root rights (like creating and installing > ports) without wracking the host system. > The users are trusted so it is not so much about security. More about keeping > the host system clean. You also could run ssh service inside the jail and give users opportunity to experiment with ssh and keys :-) ___ freebsd-stable@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/freebsd-stable To unsubscribe, send any mail to "freebsd-stable-unsubscr...@freebsd.org"
Re: jexec as user?
Thanks for all the advice. I am indeed looking for using jail from the non-root user in the host. Jailme sounds like a good solution. My use case is providing a relatively save way of giving a user the possibility to experiment with root rights (like creating and installing ports) without wracking the host system. The users are trusted so it is not so much about security. More about keeping the host system clean. Regards, Ronald. Van: Miroslav Lachman <000.f...@quip.cz> Datum: dinsdag, 19 november 2019 20:31 Aan: Christos Chatzaras , freebsd-stable CC: Ronald Klop Onderwerp: Re: jexec as user? Christos Chatzaras wrote on 2019/11/19 14:09: > > >> On 19 Nov 2019, at 15:02, mike tancsa wrote: >> >> On 11/19/2019 6:42 AM, Ronald Klop wrote: >>> Hi, >>> >>> Is it possible to jexec into a jail as a regular user. Or to enable >>> that somewhere? >>> Or is the way to do such a thing to set up ssh in the jail? >>> >> On 11.3 at least, does not the built in functionality of jexec do what >> you need ? >> >> jexec [-l] [-u username | -U username] jail [command ...] >> >> # jexec -U testuser 3 csh >> testuser@cacticonsole:/ % id >> uid=1005(testuser) gid=1005(testuser) groups=1005(testuser) >> testuser@cacticonsole:/ % >> > > I think he wants to use jexec as a normal user from the main OS. > > If he wants to run jexec as root and login to jail as user then your command works. If you want to use jexec as normal user in host, look at sysutils/jailme from ports: https://www.freshports.org/sysutils/jailme/ This version is installed setuid and does some sanity checking to ensure the username and UID match between the jail and the host system. WWW: https://github.com/Intermedix/jailme Miroslav Lachman PS: I never used jailme personally ___ freebsd-stable@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/freebsd-stable To unsubscribe, send any mail to "freebsd-stable-unsubscr...@freebsd.org"
