sigh, well that illustrates why this design was a bad idea.
I changed the param to use their short form which should keep the
service filename length under the limit.
To be totally clear, this is in no way a correct fix, I just want to
confirm this is actually the problem; then it'll have to get
Dan, I've attached /var/log/syslog and journalctl logs of a recreate
after installing nvme-cli_1.9-1ubuntu0.1+bug1874270v20210408b2_amd64 and
rebooting the host.
Apr 8 14:44:00 ICTM1608S01H1 root: JD: Resetting controller B
Apr 8 14:44:09 ICTM1608S01H1 kernel: [ 196.190003] lpfc :af:00.1:
> It looks like connect-all didn't recognize the "matching" flag
ah sorry, that param is used in later versions, i uploaded a new buidl
to the ppa can you try that one? It just started building now, so might
take some time to finish building and get published.
** Changed in: nvme-cli (Ubuntu)
Dan, I've attached /var/log/syslog and journalctl logs of a recreate
after installing nvme-cli_1.9-1ubuntu0.1+bug1874270v20210408b1_amd64 and
rebooting the host. It looks like connect-all didn't recognize the
"matching" flag.
Apr 8 11:48:45 ICTM1608S01H1 root: JD: Resetting controller B
Apr 8
@jduong can you test with the nvme-cli package from this ppa:
https://launchpad.net/~ddstreet/+archive/ubuntu/lp1874270
--
You received this bug notification because you are a member of Ubuntu
Bugs, which is subscribed to Ubuntu.
https://bugs.launchpad.net/bugs/1874270
Title:
NVMe/FC
Hmm, I looked at the udev rule and the service it calls a bit closer,
and I'm not sure if that udev rule has *ever* worked; it's attempting to
pass a full list of parameters as the template value on the cmdline, and
I can kind of understand the intention, but it's not a good way to
implement it.
Dan, where do I change the kernel rport timeout? And how can I go about
changing the timeout on a server with Emulex cards installed versus
Qlogic?
--
You received this bug notification because you are a member of Ubuntu
Bugs, which is subscribed to Ubuntu.
It looks like the service is failing because your controller is in the
process of resetting, which appears to take several minutes. I'm not
sure what the design is for nvme-cli tools handling such a long reset
time, but my first guess would be to increase the kernel rport timeout,
which appears to
At the time in which a storage controller is failed, /var/log/syslog and
journalctl look identical:
Apr 7 11:45:28 ICTM1608S01H1 kernel: [586649.657080] lpfc :af:00.1:
5:(0):6172 NVME rescanned DID x3d0a00 port_state x2
Apr 7 11:45:28 ICTM1608S01H1 kernel: [586649.657268] lpfc
> Dan, where can I find the location of these logs?
they should be in /var/log/syslog, and/or the journal, which you can
check with:
for all logs:
$ journalctl
just logs for this boot:
$ journalctl -b
just logs for the nvmf-connect unit(s):
$ journalctl -u 'nvmf-connect*'
You probably only
Dan, where can I find the location of these logs?
--
You received this bug notification because you are a member of Ubuntu
Bugs, which is subscribed to Ubuntu.
https://bugs.launchpad.net/bugs/1874270
Title:
NVMe/FC connections fail to reestablish after controller is reset
To manage
** Changed in: nvme-cli (Ubuntu)
Status: Expired => Incomplete
--
You received this bug notification because you are a member of Ubuntu
Bugs, which is subscribed to Ubuntu.
https://bugs.launchpad.net/bugs/1874270
Title:
NVMe/FC connections fail to reestablish after controller is reset
[Expired for nvme-cli (Ubuntu) because there has been no activity for 60
days.]
** Changed in: nvme-cli (Ubuntu)
Status: Incomplete => Expired
--
You received this bug notification because you are a member of Ubuntu
Bugs, which is subscribed to Ubuntu.
You'll need to check the specific output/log of the nvmf-connect@
services that are failing, e.g.:
Apr 21 16:48:53 ICTM1610S01H2 systemd-udevd[2946]: fc_udev_device:
Process 'systemctl --no-block start nvmf-connect@--device=none\t--
Status changed to 'Confirmed' because the bug affects multiple users.
** Changed in: nvme-cli (Ubuntu)
Status: New => Confirmed
--
You received this bug notification because you are a member of Ubuntu
Bugs, which is subscribed to Ubuntu.
https://bugs.launchpad.net/bugs/1874270
Title:
I am still seeing this with Ubuntu 20.04 LTS
--
You received this bug notification because you are a member of Ubuntu
Bugs, which is subscribed to Ubuntu.
https://bugs.launchpad.net/bugs/1874270
Title:
NVMe/FC connections fail to reestablish after controller is reset
To manage notifications
Also, it does not look like Broadcom's website has an autoconnect script
that supports Ubuntu.
--
You received this bug notification because you are a member of Ubuntu
Bugs, which is subscribed to Ubuntu.
https://bugs.launchpad.net/bugs/1874270
Title:
NVMe/FC connections fail to reestablish
I've attached logs.
** Attachment added: "ICTM1610S01-messages-4-21-20.zip"
https://bugs.launchpad.net/ubuntu/+source/nvme-cli/+bug/1874270/+attachment/5357988/+files/ICTM1610S01-messages-4-21-20.zip
--
You received this bug notification because you are a member of Ubuntu
Bugs, which is
18 matches
Mail list logo
