July 2016 - Neurosis - mailman3.cels.anl.gov

Re: [Neurosis] shutting neurosis down now
by Brian R. Toonen 27 Jul '16

27 Jul '16

Neurosis is back. Unfortunately, the additional IB cards couldn't be installed. The cards we had available to us are 16x and all of the slots in neurosis are 8x. --brian On 2016-07-27 11:29, Brian R. Toonen wrote: > I am shutting neurosis down now so that Bill can install some additional IB cards. I will see email again when it is back up. > > --brian >

1 0

shutting neurosis down now
by Brian R. Toonen 27 Jul '16

27 Jul '16

I am shutting neurosis down now so that Bill can install some additional IB cards. I will see email again when it is back up. --brian

1 0

Re: [Neurosis] Installing kdsa packages on Neurosis
by Zack Cornelius 07 Jul '16

07 Jul '16

The packages are installed and working. If you have any issues with kdsa, let Brian know. --Zack > From: "Zack Cornelius" <zack.cornelius(a)kove.net> > To: "neurosis" <neurosis(a)lists.jlse.anl.gov> > Sent: Thursday, July 7, 2016 10:07:05 AM > Subject: [Neurosis] Installing kdsa packages on Neurosis > All, > I'll be installing updated kdsa packages on Neurosis in about 10 minutes. I > don't expect to need to reboot the system, or interfere with any existing > processes. > I'll let you all know once the installation is complete. > --Zack > _______________________________________________ > Neurosis mailing list > Neurosis(a)lists.jlse.anl.gov > https://lists.jlse.anl.gov/mailman/listinfo/neurosis

1 0

Installing kdsa packages on Neurosis
by Zack Cornelius 07 Jul '16

07 Jul '16

All, I'll be installing updated kdsa packages on Neurosis in about 10 minutes. I don't expect to need to reboot the system, or interfere with any existing processes. I'll let you all know once the installation is complete. --Zack

1 0

Re: [Neurosis] Upgrading OFED
by Brian R. Toonen 06 Jul '16

06 Jul '16

Neurosis is back. The Kove XMem/BD driver is not installed, so the Kove malloc library and mmap'ing /xmem entries will not work right now. We will install the new driver soon. The KDSA C API, however, should be fully operational. If you encounter any problems, please let me know. --brian On 2016-07-06 15:25, Brian R. Toonen wrote: > In order to prepare for a driver update from Kove that will support Mellanox OFED 3.3, I need to update the kernel. Since no one is currently using neurosis, I'm going to do that now in addition to updating other packages with pending updates. This shouldn't take more than an hour. I will let you know when the process is complete. > > --brian > > On 2016-07-05 13:33, Allen, Benjamin S. wrote: >> This is complete. Including an updated firmware for the Mellanox HCA. Note this was a swap from RHEL's OFED to Mellanox's OFED 3.3-1.0.0.0. The latest firmware that I could find for this HP branded Mellanox ConnectX-2 HCA is 2.9.1530. >> >> I also took this chance to run JLSE's full set of Ansible playbooks over Neurosis which brought it closer to what our current configurations are. >> >> Some notables: overwrote /etc/yum.repos.d, sudoers, /etc/security/access.conf. For now I have sudoers and access.conf to allow anyone in the RAN group. It installed cpupower and get the governor to performance. Fixed rsyslog to forward logs. Updated a number of configuration to point at ftsn2 and ftsn3 instead of ftsn1 and ftsn2. Installed a handful of useful systems applications. >> >> Let me know if there's anything that appears broken. >> >> Ben >> >>> On Jul 5, 2016, at 12:49 PM, Allen, Benjamin S. <bsallen(a)alcf.anl.gov> wrote: >>> >>> Brian asked I installed / upgrade Mellanox OFED on Neurosis. Doing this now. >>> >>> Ben >>> >> >> _______________________________________________ >> Neurosis mailing list >> Neurosis(a)lists.jlse.anl.gov >> https://lists.jlse.anl.gov/mailman/listinfo/neurosis >>

1 0

Re: [Neurosis] Upgrading OFED
by Brian R. Toonen 06 Jul '16

06 Jul '16

In order to prepare for a driver update from Kove that will support Mellanox OFED 3.3, I need to update the kernel. Since no one is currently using neurosis, I'm going to do that now in addition to updating other packages with pending updates. This shouldn't take more than an hour. I will let you know when the process is complete. --brian On 2016-07-05 13:33, Allen, Benjamin S. wrote: > This is complete. Including an updated firmware for the Mellanox HCA. Note this was a swap from RHEL's OFED to Mellanox's OFED 3.3-1.0.0.0. The latest firmware that I could find for this HP branded Mellanox ConnectX-2 HCA is 2.9.1530. > > I also took this chance to run JLSE's full set of Ansible playbooks over Neurosis which brought it closer to what our current configurations are. > > Some notables: overwrote /etc/yum.repos.d, sudoers, /etc/security/access.conf. For now I have sudoers and access.conf to allow anyone in the RAN group. It installed cpupower and get the governor to performance. Fixed rsyslog to forward logs. Updated a number of configuration to point at ftsn2 and ftsn3 instead of ftsn1 and ftsn2. Installed a handful of useful systems applications. > > Let me know if there's anything that appears broken. > > Ben > >> On Jul 5, 2016, at 12:49 PM, Allen, Benjamin S. <bsallen(a)alcf.anl.gov> wrote: >> >> Brian asked I installed / upgrade Mellanox OFED on Neurosis. Doing this now. >> >> Ben >> > > _______________________________________________ > Neurosis mailing list > Neurosis(a)lists.jlse.anl.gov > https://lists.jlse.anl.gov/mailman/listinfo/neurosis >

1 0

Re: [Neurosis] Upgrading OFED
by Allen, Benjamin S. 05 Jul '16

05 Jul '16

Done. Ben > On Jul 5, 2016, at 3:37 PM, Brian R. Toonen <toonen(a)alcf.anl.gov> wrote: > > BTW, there is a bug in the Mellanox drivers version 3.1 and above. The bug only is affected when both ports are wired and configured for IB. I've temporarily disabled IB traffic over the second port by setting the port to ethernet (using connectx_port_config), although it would probably be better if we pull the cable on the second port until Mellanox fixes the bug. > > --brian > > On 2016-07-05 13:33, Allen, Benjamin S. wrote: >> This is complete. Including an updated firmware for the Mellanox HCA. Note this was a swap from RHEL's OFED to Mellanox's OFED 3.3-1.0.0.0. The latest firmware that I could find for this HP branded Mellanox ConnectX-2 HCA is 2.9.1530. >> >> I also took this chance to run JLSE's full set of Ansible playbooks over Neurosis which brought it closer to what our current configurations are. >> >> Some notables: overwrote /etc/yum.repos.d, sudoers, /etc/security/access.conf. For now I have sudoers and access.conf to allow anyone in the RAN group. It installed cpupower and get the governor to performance. Fixed rsyslog to forward logs. Updated a number of configuration to point at ftsn2 and ftsn3 instead of ftsn1 and ftsn2. Installed a handful of useful systems applications. >> >> Let me know if there's anything that appears broken. >> >> Ben >> >>> On Jul 5, 2016, at 12:49 PM, Allen, Benjamin S. <bsallen(a)alcf.anl.gov> wrote: >>> >>> Brian asked I installed / upgrade Mellanox OFED on Neurosis. Doing this now. >>> >>> Ben >>> >> >> _______________________________________________ >> Neurosis mailing list >> Neurosis(a)lists.jlse.anl.gov >> https://lists.jlse.anl.gov/mailman/listinfo/neurosis >>

1 0

Re: [Neurosis] Upgrading OFED
by Brian R. Toonen 05 Jul '16

05 Jul '16

BTW, there is a bug in the Mellanox drivers version 3.1 and above. The bug only is affected when both ports are wired and configured for IB. I've temporarily disabled IB traffic over the second port by setting the port to ethernet (using connectx_port_config), although it would probably be better if we pull the cable on the second port until Mellanox fixes the bug. --brian On 2016-07-05 13:33, Allen, Benjamin S. wrote: > This is complete. Including an updated firmware for the Mellanox HCA. Note this was a swap from RHEL's OFED to Mellanox's OFED 3.3-1.0.0.0. The latest firmware that I could find for this HP branded Mellanox ConnectX-2 HCA is 2.9.1530. > > I also took this chance to run JLSE's full set of Ansible playbooks over Neurosis which brought it closer to what our current configurations are. > > Some notables: overwrote /etc/yum.repos.d, sudoers, /etc/security/access.conf. For now I have sudoers and access.conf to allow anyone in the RAN group. It installed cpupower and get the governor to performance. Fixed rsyslog to forward logs. Updated a number of configuration to point at ftsn2 and ftsn3 instead of ftsn1 and ftsn2. Installed a handful of useful systems applications. > > Let me know if there's anything that appears broken. > > Ben > >> On Jul 5, 2016, at 12:49 PM, Allen, Benjamin S. <bsallen(a)alcf.anl.gov> wrote: >> >> Brian asked I installed / upgrade Mellanox OFED on Neurosis. Doing this now. >> >> Ben >> > > _______________________________________________ > Neurosis mailing list > Neurosis(a)lists.jlse.anl.gov > https://lists.jlse.anl.gov/mailman/listinfo/neurosis >

1 0

Re: [Neurosis] Current Sudoers
by Allen, Benjamin S. 05 Jul '16

05 Jul '16

K changed it to this list. Ben > On Jul 5, 2016, at 3:14 PM, Brian R. Toonen <toonen(a)alcf.anl.gov> wrote: > > No one outside of Bill, Kove and myself should need sudo at this point. We have setuid scripts for RAN commands that require root. I suggest a membership of allcock, toonen, zcorneli, and poling. > > --brian > > On 2016-07-05 14:04, William E. Allcock wrote: >> %RAN is fine. >> >> Bill >> -------------------------------------------------------------- >> William (Bill) E. Allcock >> allcock(a)anl.gov >> Office 1127, Bldg 240 >> 9700 S. Cass Ave >> Argonne, IL 60439 >> Office: 630-252-7573 >> Mobile: 630-854-2842 >> Fax: 630-252-6585 >> >> >> The test of courage comes when we are in the minority. >> The test of tolerance comes when we are in the majority. >> - Ralph W. Sockman >> >>> On Jul 5, 2016, at 1:53 PM, Allen, Benjamin S. <bsallen(a)alcf.anl.gov> wrote: >>> >>> I'm fine with leaving sudoers access to be all of %RAN. Although if you want me to reduce it to Brian and you, that's easy todo as well. >>> >>> Ben >>> >>>> On Jul 5, 2016, at 1:38 PM, William E. Allcock <allcock(a)anl.gov> wrote: >>>> >>>> Ben, >>>> >>>> I see from your follow on mail that for now you have it set to anyone in the RAN group, which I think is fine for now. >>>> >>>> Brian is off today, so he probably wont chime in until tomorrow, but if you want to trim the sudoers list down, it could probably be just Brian and I, plus systems people. I think we have permissions and scripts to handle things people needed sudo for in the past, but Brian will have to confirm that. >>>> >>>> Bill >>>> -------------------------------------------------------------- >>>> William (Bill) E. Allcock >>>> allcock(a)anl.gov >>>> Office 1127, Bldg 240 >>>> 9700 S. Cass Ave >>>> Argonne, IL 60439 >>>> Office: 630-252-7573 >>>> Mobile: 630-854-2842 >>>> Fax: 630-252-6585 >>>> >>>> >>>> The test of courage comes when we are in the minority. >>>> The test of tolerance comes when we are in the majority. >>>> - Ralph W. Sockman >>>> >>>>> On Jul 5, 2016, at 1:09 PM, Allen, Benjamin S. <bsallen(a)alcf.anl.gov> wrote: >>>>> >>>>> Bill and Brian, >>>>> >>>>> Who all should have sudo access on Neurosis? >>>>> >>>>> Thanks, >>>>> >>>>> Ben >>>> >>> >>

1 0

Re: [Neurosis] Current Sudoers
by Brian R. Toonen 05 Jul '16

05 Jul '16

No one outside of Bill, Kove and myself should need sudo at this point. We have setuid scripts for RAN commands that require root. I suggest a membership of allcock, toonen, zcorneli, and poling. --brian On 2016-07-05 14:04, William E. Allcock wrote: > %RAN is fine. > > Bill > -------------------------------------------------------------- > William (Bill) E. Allcock > allcock(a)anl.gov > Office 1127, Bldg 240 > 9700 S. Cass Ave > Argonne, IL 60439 > Office: 630-252-7573 > Mobile: 630-854-2842 > Fax: 630-252-6585 > > > The test of courage comes when we are in the minority. > The test of tolerance comes when we are in the majority. > - Ralph W. Sockman > >> On Jul 5, 2016, at 1:53 PM, Allen, Benjamin S. <bsallen(a)alcf.anl.gov> wrote: >> >> I'm fine with leaving sudoers access to be all of %RAN. Although if you want me to reduce it to Brian and you, that's easy todo as well. >> >> Ben >> >>> On Jul 5, 2016, at 1:38 PM, William E. Allcock <allcock(a)anl.gov> wrote: >>> >>> Ben, >>> >>> I see from your follow on mail that for now you have it set to anyone in the RAN group, which I think is fine for now. >>> >>> Brian is off today, so he probably wont chime in until tomorrow, but if you want to trim the sudoers list down, it could probably be just Brian and I, plus systems people. I think we have permissions and scripts to handle things people needed sudo for in the past, but Brian will have to confirm that. >>> >>> Bill >>> -------------------------------------------------------------- >>> William (Bill) E. Allcock >>> allcock(a)anl.gov >>> Office 1127, Bldg 240 >>> 9700 S. Cass Ave >>> Argonne, IL 60439 >>> Office: 630-252-7573 >>> Mobile: 630-854-2842 >>> Fax: 630-252-6585 >>> >>> >>> The test of courage comes when we are in the minority. >>> The test of tolerance comes when we are in the majority. >>> - Ralph W. Sockman >>> >>>> On Jul 5, 2016, at 1:09 PM, Allen, Benjamin S. <bsallen(a)alcf.anl.gov> wrote: >>>> >>>> Bill and Brian, >>>> >>>> Who all should have sudo access on Neurosis? >>>> >>>> Thanks, >>>> >>>> Ben >>> >> >

1 0