Re: [Freeipmi-users] Free ipmi 1.6.3 ipmi-sel command exits with a error on some quanta servers
That works perfectly on the quanta, thank you so much. Let me build it for a few different OSes and make sure it works on all our other hardware. --Stephen On 8/15/19, 9:20 AM, "Albert Chu" wrote: External email: Use caution opening links or attachments Hi Stephen, I thought of a really simple fix. I have a branch called "event_only_record_column_widths" on github here: https://github.com/chu11/freeipmi-mirror/tree/event_only_record_column_widths Can you see if this works for you? Al On Wed, 2019-08-14 at 21:06 -0700, Albert Chu wrote: > Hi Stephen, > > > 1) Using "--legacy-output" works great. That is an excellent work > > around. Do you have any idea what information it is lacking? > > It's mostly details. Like sensor values will not output on threshold > events, event states aren't available, and various OEM support will > be > missing. And probably a lot of other things like that. > > I'll try to get a temporary fix out sometime this week. > > Al > > On Wed, Aug 14, 2019 at 5:08 PM Stephen Abbene > wrote: > > > > Hello Albert, > > > > 1) Using "--legacy-output" works great. That is an excellent work > > around. Do you have any idea what information it is lacking? > > 2) Using "--comma-separated-output" does not work. I end up > > getting the same, "ipmi_sdr_parse_sensor_units: invalid sdr record > > ", error. > > $ sudo ./ipmi-sel --comma-separated-output > > Caching SDR repository information: /root/.freeipmi/sdr-cache/sdr- > > cache-.localhost > > Caching SDR record 113 of 113 (current record ID 113) > > ipmi_sdr_parse_sensor_units: invalid sdr record > > > > We have opened a case with our vendor based on your findings I will > > let you know if they follow up. This is really helpful for us. We > > use freeipmi in many of our critical monitoring applications and > > really appreciate all the work you put into it. > > > > Thanks, > > Stephen Abbene > > > > On 8/14/19, 4:19 PM, "Albert Chu" wrote: > > > > External email: Use caution opening links or attachments > > > > > > Hi Stephen, > > > > While skimming code, it occurred to me that if we don't bother > > to > > calculate column widths for pretty output, ipmi-sel might work > > for you. > > > > 1) does "ipmi-sel --legacy-output" work > > > > the output is ugly and not as useful as the pretty output > > (lacks > > information from newer releases), but it's something for you > > temporarily > > > > 2) does "ipmi-sel --comma-separated-output" > > > > skimming through code, I think I may stupidly calculate columns > > widths > > even though the user requested comma separated output. So > > maybe this > > won't work and is something I need to fix. > > > > If #1 works, then that gives us a relatively simple avenue for > > a fix. > > If sensor units aren't available, just don't use that record's > > sensor > > units in column width calculations. > > > > There could be additional problems if the vendor actually > > outputs a > > real event that needs information from that SDR record. But I > > would > > consider that unlikely. > > > > Al > > > > On Wed, 2019-08-14 at 22:07 +, Stephen Abbene wrote: > > > Thank you for your analysis Albert. I will reach out to the > > vendor > > > with your observations.Here is the additional debugging > > output > > > you requested. The first is the output of ipmi-sel without > > > discretereading, and the second is a build of freeipmi from > > master > > > branch on git run with ipmi-sel --debug. > > -- > > Albert Chu > > ch...@llnl.gov > > Computer Scientist > > High Performance Systems Division > > Lawrence Livermore National Laboratory > > > > > > > > > > ___ > > Freeipmi-users mailing list > > Freeipmi-users@gnu.org > > https://lists.gnu.org/mailman/listinfo/freeipmi-users -- Albert Chu ch...@llnl.gov Computer Scientist High Performance Systems Division Lawrence Livermore National Laboratory ___ Freeipmi-users mailing list Freeipmi-users@gnu.org https://lists.gnu.org/mailman/listinfo/freeipmi-users
Re: [Freeipmi-users] Free ipmi 1.6.3 ipmi-sel command exits with a error on some quanta servers
Hi Stephen, I thought of a really simple fix. I have a branch called "event_only_record_column_widths" on github here: https://github.com/chu11/freeipmi-mirror/tree/event_only_record_column_widths Can you see if this works for you? Al On Wed, 2019-08-14 at 21:06 -0700, Albert Chu wrote: > Hi Stephen, > > > 1) Using "--legacy-output" works great. That is an excellent work > > around. Do you have any idea what information it is lacking? > > It's mostly details. Like sensor values will not output on threshold > events, event states aren't available, and various OEM support will > be > missing. And probably a lot of other things like that. > > I'll try to get a temporary fix out sometime this week. > > Al > > On Wed, Aug 14, 2019 at 5:08 PM Stephen Abbene > wrote: > > > > Hello Albert, > > > > 1) Using "--legacy-output" works great. That is an excellent work > > around. Do you have any idea what information it is lacking? > > 2) Using "--comma-separated-output" does not work. I end up > > getting the same, "ipmi_sdr_parse_sensor_units: invalid sdr record > > ", error. > > $ sudo ./ipmi-sel --comma-separated-output > > Caching SDR repository information: /root/.freeipmi/sdr-cache/sdr- > > cache-.localhost > > Caching SDR record 113 of 113 (current record ID 113) > > ipmi_sdr_parse_sensor_units: invalid sdr record > > > > We have opened a case with our vendor based on your findings I will > > let you know if they follow up. This is really helpful for us. We > > use freeipmi in many of our critical monitoring applications and > > really appreciate all the work you put into it. > > > > Thanks, > > Stephen Abbene > > > > On 8/14/19, 4:19 PM, "Albert Chu" wrote: > > > > External email: Use caution opening links or attachments > > > > > > Hi Stephen, > > > > While skimming code, it occurred to me that if we don't bother > > to > > calculate column widths for pretty output, ipmi-sel might work > > for you. > > > > 1) does "ipmi-sel --legacy-output" work > > > > the output is ugly and not as useful as the pretty output > > (lacks > > information from newer releases), but it's something for you > > temporarily > > > > 2) does "ipmi-sel --comma-separated-output" > > > > skimming through code, I think I may stupidly calculate columns > > widths > > even though the user requested comma separated output. So > > maybe this > > won't work and is something I need to fix. > > > > If #1 works, then that gives us a relatively simple avenue for > > a fix. > > If sensor units aren't available, just don't use that record's > > sensor > > units in column width calculations. > > > > There could be additional problems if the vendor actually > > outputs a > > real event that needs information from that SDR record. But I > > would > > consider that unlikely. > > > > Al > > > > On Wed, 2019-08-14 at 22:07 +, Stephen Abbene wrote: > > > Thank you for your analysis Albert. I will reach out to the > > vendor > > > with your observations.Here is the additional debugging > > output > > > you requested. The first is the output of ipmi-sel without > > > discretereading, and the second is a build of freeipmi from > > master > > > branch on git run with ipmi-sel --debug. > > -- > > Albert Chu > > ch...@llnl.gov > > Computer Scientist > > High Performance Systems Division > > Lawrence Livermore National Laboratory > > > > > > > > > > ___ > > Freeipmi-users mailing list > > Freeipmi-users@gnu.org > > https://lists.gnu.org/mailman/listinfo/freeipmi-users -- Albert Chu ch...@llnl.gov Computer Scientist High Performance Systems Division Lawrence Livermore National Laboratory ___ Freeipmi-users mailing list Freeipmi-users@gnu.org https://lists.gnu.org/mailman/listinfo/freeipmi-users
Re: [Freeipmi-users] Free ipmi 1.6.3 ipmi-sel command exits with a error on some quanta servers
Hi Stephen, > 1) Using "--legacy-output" works great. That is an excellent work around. > Do you have any idea what information it is lacking? It's mostly details. Like sensor values will not output on threshold events, event states aren't available, and various OEM support will be missing. And probably a lot of other things like that. I'll try to get a temporary fix out sometime this week. Al On Wed, Aug 14, 2019 at 5:08 PM Stephen Abbene wrote: > > Hello Albert, > > 1) Using "--legacy-output" works great. That is an excellent work around. > Do you have any idea what information it is lacking? > 2) Using "--comma-separated-output" does not work. I end up getting the > same, "ipmi_sdr_parse_sensor_units: invalid sdr record ", error. > $ sudo ./ipmi-sel --comma-separated-output > Caching SDR repository information: > /root/.freeipmi/sdr-cache/sdr-cache-.localhost > Caching SDR record 113 of 113 (current record ID 113) > ipmi_sdr_parse_sensor_units: invalid sdr record > > We have opened a case with our vendor based on your findings I will let you > know if they follow up. This is really helpful for us. We use freeipmi in > many of our critical monitoring applications and really appreciate all the > work you put into it. > > Thanks, > Stephen Abbene > > On 8/14/19, 4:19 PM, "Albert Chu" wrote: > > External email: Use caution opening links or attachments > > > Hi Stephen, > > While skimming code, it occurred to me that if we don't bother to > calculate column widths for pretty output, ipmi-sel might work for you. > > 1) does "ipmi-sel --legacy-output" work > > the output is ugly and not as useful as the pretty output (lacks > information from newer releases), but it's something for you > temporarily > > 2) does "ipmi-sel --comma-separated-output" > > skimming through code, I think I may stupidly calculate columns widths > even though the user requested comma separated output. So maybe this > won't work and is something I need to fix. > > If #1 works, then that gives us a relatively simple avenue for a fix. > If sensor units aren't available, just don't use that record's sensor > units in column width calculations. > > There could be additional problems if the vendor actually outputs a > real event that needs information from that SDR record. But I would > consider that unlikely. > > Al > > On Wed, 2019-08-14 at 22:07 +, Stephen Abbene wrote: > > Thank you for your analysis Albert. I will reach out to the vendor > > with your observations.Here is the additional debugging output > > you requested. The first is the output of ipmi-sel without > > discretereading, and the second is a build of freeipmi from master > > branch on git run with ipmi-sel --debug. > -- > Albert Chu > ch...@llnl.gov > Computer Scientist > High Performance Systems Division > Lawrence Livermore National Laboratory > > > > > ___ > Freeipmi-users mailing list > Freeipmi-users@gnu.org > https://lists.gnu.org/mailman/listinfo/freeipmi-users ___ Freeipmi-users mailing list Freeipmi-users@gnu.org https://lists.gnu.org/mailman/listinfo/freeipmi-users
Re: [Freeipmi-users] Free ipmi 1.6.3 ipmi-sel command exits with a error on some quanta servers
Hello Albert, 1) Using "--legacy-output" works great. That is an excellent work around. Do you have any idea what information it is lacking? 2) Using "--comma-separated-output" does not work. I end up getting the same, "ipmi_sdr_parse_sensor_units: invalid sdr record ", error. $ sudo ./ipmi-sel --comma-separated-output Caching SDR repository information: /root/.freeipmi/sdr-cache/sdr-cache-.localhost Caching SDR record 113 of 113 (current record ID 113) ipmi_sdr_parse_sensor_units: invalid sdr record We have opened a case with our vendor based on your findings I will let you know if they follow up. This is really helpful for us. We use freeipmi in many of our critical monitoring applications and really appreciate all the work you put into it. Thanks, Stephen Abbene On 8/14/19, 4:19 PM, "Albert Chu" wrote: External email: Use caution opening links or attachments Hi Stephen, While skimming code, it occurred to me that if we don't bother to calculate column widths for pretty output, ipmi-sel might work for you. 1) does "ipmi-sel --legacy-output" work the output is ugly and not as useful as the pretty output (lacks information from newer releases), but it's something for you temporarily 2) does "ipmi-sel --comma-separated-output" skimming through code, I think I may stupidly calculate columns widths even though the user requested comma separated output. So maybe this won't work and is something I need to fix. If #1 works, then that gives us a relatively simple avenue for a fix. If sensor units aren't available, just don't use that record's sensor units in column width calculations. There could be additional problems if the vendor actually outputs a real event that needs information from that SDR record. But I would consider that unlikely. Al On Wed, 2019-08-14 at 22:07 +, Stephen Abbene wrote: > Thank you for your analysis Albert. I will reach out to the vendor > with your observations.Here is the additional debugging output > you requested. The first is the output of ipmi-sel without > discretereading, and the second is a build of freeipmi from master > branch on git run with ipmi-sel --debug. -- Albert Chu ch...@llnl.gov Computer Scientist High Performance Systems Division Lawrence Livermore National Laboratory ___ Freeipmi-users mailing list Freeipmi-users@gnu.org https://lists.gnu.org/mailman/listinfo/freeipmi-users
Re: [Freeipmi-users] Free ipmi 1.6.3 ipmi-sel command exits with a error on some quanta servers
Hi Stephen, While skimming code, it occurred to me that if we don't bother to calculate column widths for pretty output, ipmi-sel might work for you. 1) does "ipmi-sel --legacy-output" work the output is ugly and not as useful as the pretty output (lacks information from newer releases), but it's something for you temporarily 2) does "ipmi-sel --comma-separated-output" skimming through code, I think I may stupidly calculate columns widths even though the user requested comma separated output. So maybe this won't work and is something I need to fix. If #1 works, then that gives us a relatively simple avenue for a fix. If sensor units aren't available, just don't use that record's sensor units in column width calculations. There could be additional problems if the vendor actually outputs a real event that needs information from that SDR record. But I would consider that unlikely. Al On Wed, 2019-08-14 at 22:07 +, Stephen Abbene wrote: > Thank you for your analysis Albert. I will reach out to the vendor > with your observations.Here is the additional debugging output > you requested. The first is the output of ipmi-sel without > discretereading, and the second is a build of freeipmi from master > branch on git run with ipmi-sel --debug. -- Albert Chu ch...@llnl.gov Computer Scientist High Performance Systems Division Lawrence Livermore National Laboratory ___ Freeipmi-users mailing list Freeipmi-users@gnu.org https://lists.gnu.org/mailman/listinfo/freeipmi-users
Re: [Freeipmi-users] Free ipmi 1.6.3 ipmi-sel command exits with a error on some quanta servers
Hi Stephen, So on the very last record we see [ 1h] = event_reading_type_code[ 8b] which means a threshold based sensor, which typically means a temperature, rpm or something similar sensor. And we see: [ Bh] = sensor_type[ 8b] Which is a "other units based sensor". So given its a threshold sensor and the type is "other units", there are units within this SDR record. But there is not. So I'm thinking it's an error on the manufacturer's part. But clearly we should not just exit on this error and should have ipmi-sel work around it in some way, like output "SEL record un-parseable" or something and have ipmi-sel output the rest of the SEL. I have to think about how to do this for a bit. In the mean time, would you be willing to re-run without the "discretereading" flag. I just want to double check that the same record shows the same problem. You don't have to send output. Just make sure that the same event at the end of the output occurs. And perhaps make sure the problem occurs under the master branch of FreeIPMI, b/c I'll probably want to have you try a patch at some point :-) Thanks, Al On Wed, 2019-08-14 at 20:39 +, Stephen Abbene wrote: > = > SDR Event Only Record > = > [ 3Fh] = record_id[16b] > [ 1h] = sdr_version_major[ 4b] > [ 5h] = sdr_version_minor[ 4b] > [ 3h] = record_type[ 8b] > [ 19h] = record_length[ 8b] > [ 0h] = sensor_owner_id.type[ 1b] > [ 16h] = sensor_owner_id[ 7b] > [ 0h] = sensor_owner_lun[ 2b] > [ 0h] = fru_inventory_device_owner_lun[ 2b] > [ 6h] = channel_number[ 4b] > [ ADh] = sensor_number[ 8b] > [ 3h] = entity_id[ 8b] > [ 4h] = entity_instance[ 7b] > [ 0h] = entity_instance.type[ 1b] > [ Bh] = sensor_type[ 8b] > [ 1h] = event_reading_type_code[ 8b] > [ 1h] = share_count[ 4b] > [ 0h] = id_string_instance_modifier_type[ 2b] > [ 0h] = sensor_direction[ 2b] > [ 0h] = id_string_instance_modifier_offset[ 7b] > [ 0h] = entity_instance_sharing[ 1b] > [ 0h] = reserved[ 8b] > [ 0h] = oem[ 8b] > [ CDh] = id_string_type_length_code[ 8b] > [ BYTE ARRAY ... ] = id_string[13B] > [ 43h 68h 61h 73h 73h 69h 73h 20h ] > [ 50h 6Fh 77h 65h 72h ] > ipmi_sdr_parse_sensor_units: invalid sdr record -- Albert Chu ch...@llnl.gov Computer Scientist High Performance Systems Division Lawrence Livermore National Laboratory ___ Freeipmi-users mailing list Freeipmi-users@gnu.org https://lists.gnu.org/mailman/listinfo/freeipmi-users
Re: [Freeipmi-users] Free ipmi 1.6.3 ipmi-sel command exits with a error on some quanta servers
Hi Stephen, This is an error I don't recall seeing before. It appears one of the SDR records is not legal and FreeIPMI doesn't really know how to proceed b/c of it. No workaround I can think about at the moment. Could you send me the --debug output. Minimally I can see what FreeIPMI thinks this motherboard is doing. Then we consider workaround options that I might add. Al On Wed, 2019-08-14 at 19:25 +, Stephen Abbene wrote: > Hello, > > I noticed that on some quanta systems, particularly the QuantaPlex > T42D-2U systems that the ipmi-sel command exits with an error, " > ipmi_sdr_parse_sensor_units: invalid sdr record". Does anyone have > any suggestions on how to get this working? > > $ sudo dmidecode -s system-product-name > QuantaPlex T42D-2U > > $ sudo ./ipmi-sel --version > ipmi-sel - 1.6.3 > Copyright (C) 2003-2015 FreeIPMI Core Team > This program is free software; you may redistribute it under the > terms of > the GNU General Public License. This program has absolutely no > warranty. > > $ sudo rm -rf /root/.freeipmi > > $ sudo ./ipmi-sel -f > Flushing cache: /root/.freeipmi/sdr-cache/sdr-cache-.localhost > [sabbene@rno-sim-g08-196 ipmi-sel]$ sudo ./ipmi-sel -W > discretereading > Caching SDR repository information: /root/.freeipmi/sdr-cache/sdr- > cache-.localhost > Caching SDR record 113 of 113 (current record ID 113) > ipmi_sdr_parse_sensor_units: invalid sdr record > > $ sudo ./ipmi-sel -W discretereading > Caching SDR repository information: /root/.freeipmi/sdr-cache/sdr- > cache-.localhost > Caching SDR record 113 of 113 (current record ID 113) > ipmi_sdr_parse_sensor_units: invalid sdr record > > $ sudo ipmitool mc info > Device ID : 32 > Device Revision : 1 > Firmware Revision : 3.67 > IPMI Version : 2.0 > Manufacturer ID : 7244 > Manufacturer Name : Quanta > Product ID: 13636 (0x3544) > Product Name : Unknown (0x3544) > Device Available : yes > Provides Device SDRs : yes > Additional Device Support : > Sensor Device > SDR Repository Device > SEL Device > FRU Inventory Device > IPMB Event Receiver > IPMB Event Generator > Chassis Device > Aux Firmware Rev Info : > 0x00 > 0x00 > 0x00 > 0x00 > > > Thanks, > Stephen Abbene > > ___ > Freeipmi-users mailing list > Freeipmi-users@gnu.org > https://lists.gnu.org/mailman/listinfo/freeipmi-users -- Albert Chu ch...@llnl.gov Computer Scientist High Performance Systems Division Lawrence Livermore National Laboratory ___ Freeipmi-users mailing list Freeipmi-users@gnu.org https://lists.gnu.org/mailman/listinfo/freeipmi-users