Date: Thu, 31 Mar 2022 23:09:45 -0600 From: Warner Losh <imp@bsdimp.com> To: George Michaelson <ggm@algebras.org> Cc: FreeBSD Stable <freebsd-stable@freebsd.org> Subject: Re: CAM timeouts on boot (13.0-p8, Dell R730xd, mrsas controller) Message-ID: <CANCZdfpk3_Vit2=7E0xQ8CD1oS1Tj8xAB7vH5dh-toemCWQkqA@mail.gmail.com> In-Reply-To: <CAKr6gn05CLzYE-%2BnJScsSUqZHTZ%2B21f%2BGRUcXa%2BC8isPypV7bw@mail.gmail.com> References: <CAKr6gn1KEk5T0W6yJniKKSfHYaNdGh=BhS5LUJ=Htg2mWeTR5g@mail.gmail.com> <CANCZdfryHTfA0gTs6eEkMG9x6WxDo%2BsMijQqJuUbEdaO%2B546pA@mail.gmail.com> <CAKr6gn3dh_Os9-hWjOk_qgWdUHwjfdN1MSFQLqz-3zuHn0SegQ@mail.gmail.com> <CAKr6gn05CLzYE-%2BnJScsSUqZHTZ%2B21f%2BGRUcXa%2BC8isPypV7bw@mail.gmail.com>
next in thread | previous in thread | raw e-mail | index | archive | help
--0000000000009588ac05db90c982 Content-Type: text/plain; charset="UTF-8" On Thu, Mar 31, 2022 at 9:40 PM George Michaelson <ggm@algebras.org> wrote: > I checked /boot/zfs/zpool.cache and /etc/zfs/zpool.cache both have the > tank in question, zdb agrees, they are identical, multiple reboots > re-tested: without a manual zpool import tank, only the zfs zroot > which is magically mounted appears. The other tank does not unless you > call it into being. Doing that after multiuser/daemon is bad if your > daemons work in it. (and bad for NFS exports) > > The zfs pages are a bit cryptic about altroot and other zpool get all > values, but the basics seem to be that if you do zpool import, and its > in cache, then it should be detected during zfs/zpool activation. > People expect that once you import, it will re-import automatically > thereafter unless you tell it not to. I checked zpool values on a > system which doesnt have the problem and I don't see any variance for > zpool params. zpool.cache was the recommended "have you checked" > reference. > > Not that this can't be PBCAK, but I have tried not to be the root > cause here. It happened across my upgrade, and the CAM timeout > intruded in the same window. > I'd expect this to work, though there was a change in this file's location between pre-openzfs and now. But if you are using 12.x, then that's not going to be the case... dmesg is interesting, but not completely helpful... Do can you share your zfs config details (what disks are in what datasets?) Warner > G > > On Fri, Apr 1, 2022 at 1:31 PM George Michaelson <ggm@algebras.org> wrote: > > > > Thanks for the cluestick. I don't disagree with anything you said btw. > > > > Here's the dmesg. (attachment, can in-line if thats better) > --0000000000009588ac05db90c982 Content-Type: text/html; charset="UTF-8" Content-Transfer-Encoding: quoted-printable <div dir=3D"ltr"><div dir=3D"ltr"><br></div><br><div class=3D"gmail_quote">= <div dir=3D"ltr" class=3D"gmail_attr">On Thu, Mar 31, 2022 at 9:40 PM Georg= e Michaelson <<a href=3D"mailto:ggm@algebras.org">ggm@algebras.org</a>&g= t; wrote:<br></div><blockquote class=3D"gmail_quote" style=3D"margin:0px 0p= x 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex">I chec= ked /boot/zfs/zpool.cache and /etc/zfs/zpool.cache both have the<br> tank in question, zdb agrees, they are identical, multiple reboots<br> re-tested: without a manual zpool import tank, only the zfs zroot<br> which is magically mounted appears. The other tank does not unless you<br> call it into being. Doing that after multiuser/daemon is bad if your<br> daemons work in it. (and bad for NFS exports)<br> <br> The zfs pages are a bit cryptic about altroot and other zpool get all<br> values, but the basics seem to be that if you do zpool import, and its<br> in cache, then it should be detected during zfs/zpool activation.<br> People expect that once you import, it will re-import automatically<br> thereafter unless you tell it not to. I checked zpool values on a<br> system which doesnt have the problem and I don't see any variance for<b= r> zpool params.=C2=A0 zpool.cache was the recommended "have you checked&= quot;<br> reference.<br> <br> Not that this can't be PBCAK, but I have tried not to be the root<br> cause here. It happened across my upgrade, and the CAM timeout<br> intruded in the same window.<br></blockquote><div><br></div><div>I'd ex= pect this to work, though there was a change in this file's location be= tween</div><div>pre-openzfs and now. But if you are using 12.x, then that&#= 39;s not going to be the case...</div><div><br></div><div>dmesg is interest= ing, but not completely helpful... Do can you share your zfs config</div><d= iv>details (what disks are in what datasets?)</div><div><br></div><div>Warn= er<br></div><div>=C2=A0</div><blockquote class=3D"gmail_quote" style=3D"mar= gin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1= ex"> G<br> <br> On Fri, Apr 1, 2022 at 1:31 PM George Michaelson <<a href=3D"mailto:ggm@= algebras.org" target=3D"_blank">ggm@algebras.org</a>> wrote:<br> ><br> > Thanks for the cluestick. I don't disagree with anything you said = btw.<br> ><br> > Here's the dmesg.=C2=A0 (attachment, can in-line if thats better)<= br> </blockquote></div></div> --0000000000009588ac05db90c982--
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?CANCZdfpk3_Vit2=7E0xQ8CD1oS1Tj8xAB7vH5dh-toemCWQkqA>