Smart-UPS RT 10000 XL support with apcsmart

[surr]

I use Smart-UPS RT 10000 XL device with debian packaged nut. After upgrading from debian 6 to 7 (and nut from 2.6.4 to 2.7.2) a lot of features disappear. For example, there is only voltage value under input.* hierarchy in upsc output and upscmd -l returns an empty list. All missing things worked fine with previous version of nut. I have to recalibrate device after battery replacement, but I do not know how to do it now.

Link to device at vendor site. The device is equipped with DB-9 connector, not a RJ-45 one.

It may be irrelevant, but somewhat similar model Smart-UPS 3000 RM works as before after upgrade. This device returns a lot of data and all commands are accessible.

[clepple]

As mentioned in the NEWS file (available at /usr/share/doc/nut-server/NEWS.gz in Debian) and in the man page, we replaced the apcsmart driver with a new version. The version from NUT 2.6.4 is still available as apcsmart-old.

However, apcsmart-old will eventually be removed from NUT. Can you post the output of upsc?

Tagging @msoltyspl in case I missed something.

[surr]

Output with apcsmart driver:

battery.charge: 100.0
battery.voltage: 218.1
device.mfr: APC
device.model: Smart-UPS RT 10000 XL   
device.serial: B21130002653  
device.type: ups
driver.name: apcsmart
driver.parameter.cable: 940-0024B
driver.parameter.pollinterval: 2
driver.parameter.port: /dev/ups
driver.parameter.sdtype: 1
driver.version: 2.6.4
driver.version.internal: 3.04
input.voltage: 223.2
output.current: 07.29
output.voltage: 231.6
ups.firmware: 476.18.W
ups.load: 015.0
ups.mfr: APC
ups.model: Smart-UPS RT 10000 XL   
ups.serial: B21130002653  
ups.status: OL
ups.temperature: 024.6

Output with apcsmart-old driver:

battery.alarm.threshold: 0
battery.charge: 100.0
battery.charge.restart: 00
battery.date: 07/22/11
battery.packs: 001
battery.packs.bad: 000
battery.runtime: 3960
battery.runtime.low: 120
battery.voltage: 219.2
battery.voltage.nominal: 192
device.mfr: APC
device.model: Smart-UPS RT 10000 XL   
device.serial: B21130002653  
device.type: ups
driver.name: apcsmart-old
driver.parameter.cable: 940-0024B
driver.parameter.pollinterval: 2
driver.parameter.port: /dev/ups
driver.parameter.sdtype: 1
driver.version: 2.6.4
driver.version.internal: 2.1
input.frequency: 49.99
input.quality: FF
input.sensitivity: A
input.transfer.high: 253
input.transfer.low: 161
input.transfer.reason: simulated power failure or UPS test
input.voltage: 224.6
input.voltage.maximum: 224.6
input.voltage.minimum: 220.3
output.current: 07.29
output.voltage: 230.4
output.voltage.nominal: 230
ups.delay.shutdown: 020
ups.delay.start: 000
ups.display.language: E
ups.firmware: 476.18.W
ups.firmware.old: 5YW
ups.id: UPS_IDEN
ups.load: 015.0
ups.mfr: APC
ups.mfr.date: 07/22/11
ups.model: Smart-UPS RT 10000 XL   
ups.serial: B21130002653  
ups.status: OL
ups.temperature: 024.6
ups.test.interval: 1209600
ups.test.result: NO

[surr]

And the actual upgrade was from 2.4.3 to 2.6.4, of course. 2.7.2 is the version from the upcoming debian release.

[clepple]

You could try the following: stop the driver, then run '/lib/nut/apcsmart -D -a name-of-UPS 2>&1 | tee /tmp/apc-log.txt', then post the log in a Gist or pastebin, and post the URL here.

I think there is a firmware table that isn't matching your version.

[surr]

   0.000000 debug level is '3'
   0.076022 attempting firmware lookup using command 'V'
   0.156413 firmware: [5YW]
   0.156429 firmware not found in compatibility table - trying normal method
   0.156433 APC - attempting to find command set
   1.096035 valid_cmd: cmd [a] failed regex match
   1.096058 APC - trying to handle unknown model
   2.985510 detected Matrix-UPS [B21130002653  ] on /dev/ups
   3.055928 update_info_all: starting
   4.745038 update_info_all: done
   4.745136 dstate_init: sock /var/run/nut/apcsmart-ups open on fd 5
   4.865073 update_info_normal: starting
   5.764635 update_info_normal: done

[clepple]

On Jan 18, 2015, at 9:25 AM, surr notifications@github.com wrote:

And the actual upgrade was from 2.4.3 to 2.6.4, of course. 2.7.2 is the version from the upcoming debian release.

Oh, then that means that we had a different APC driver for 2.4.3, 2.6.4 and 2.7.2.

Can you install the 2.7.2 version from testing? I would hate to try and fix a bug in what is now the old driver.

[surr]

It is hard to install new version from debian, so I build nut from git with these parameters.

./configure --without-all --with-serial --prefix=/tmp/nut --with-statepath=/tmp/nut/_

upsc output

battery.charge: 100.0
battery.voltage: 219.2
device.mfr: APC
device.model: Smart-UPS RT 10000 XL   
device.serial: B21130002653  
device.type: ups
driver.name: apcsmart
driver.parameter.cable: 940-0024B
driver.parameter.pollinterval: 2
driver.parameter.port: /dev/ups
driver.parameter.sdtype: 1
driver.version: 2.7.2-signed-145-g26d096f
driver.version.internal: 3.1
input.voltage: 227.5
output.current: 07.29
output.voltage: 230.4
ups.firmware: 476.18.W
ups.load: 016.0
ups.mfr: APC
ups.model: Smart-UPS RT 10000 XL   
ups.serial: B21130002653  
ups.status: OL
ups.temperature: 022.9

Debug log:

Network UPS Tools - APC Smart protocol driver 3.1 (2.7.2-signed-145-g26d096f)
APC command table version 3.1
   0.000000 debug level is '1'
   0.076376 attempting firmware lookup using [V]
   0.156374 detected firmware version: 5YW
   0.156433 attempting var/cmdset lookup using [a]
   1.096819 very old or unknown APC model, support will be limited
   1.096848 preread_data: ups.temperature [C]
   1.216747 ups.temperature [C] - variable supported
   1.216761 preread_data: ups.load [P]
   1.326769 ups.load [P] - variable supported
   1.326797 preread_data: input.voltage [L]
   1.496736 input.voltage [L] - variable supported
   1.496748 preread_data: output.voltage [O]
   1.606750 output.voltage [O] - variable supported
   1.606763 preread_data: battery.charge [f]
   1.737142 battery.charge [f] - variable supported
   1.737153 preread_data: battery.voltage [B]
   1.896782 battery.voltage [B] - variable supported
   1.896813 preread_data: ups.model [0x01]
   2.207153 ups.model [0x01] - variable supported
   2.207179 preread_data: ups.serial [n]
   2.467158 ups.serial [n] - variable supported
   2.467185 preread_data: output.current [/]
   2.647135 output.current [/] - variable supported
   2.647151 preread_data: ups.firmware [b]
   2.777179 ups.firmware [b] - variable combination supported
   2.777194 detected Matrix-UPS [B21130002653  ] on /dev/ups
   2.777201 update_status: [Q]
   2.897141 update_info: starting scan (all vars)
   2.897161 poll_data: ups.temperature [C]
   3.007150 poll_data: ups.load [P]
   3.117135 poll_data: input.voltage [L]
   3.237123 poll_data: output.current [/]
   3.347127 poll_data: output.voltage [O]
   3.517558 poll_data: battery.charge [f]
   3.627115 poll_data: battery.voltage [B]
   3.797106 poll_data: ups.serial [n]
   4.007514 poll_data: ups.model [0x01]
   4.367490 poll_data: ups.firmware [b]
   4.547491 update_info: scan completed
   4.547585 update_status: [Q]
   4.617503 update_info: starting scan
   4.617517 poll_data: ups.temperature [C]
   4.767515 poll_data: ups.load [P]
   4.877499 poll_data: input.voltage [L]
   5.047491 poll_data: output.current [/]
   5.157508 poll_data: output.voltage [O]
   5.327496 poll_data: battery.charge [f]
   5.437505 poll_data: battery.voltage [B]
   5.607506 update_info: scan completed

[clepple]

0.076376 attempting firmware lookup using [V] 0.156374 detected firmware version: 5YW 0.156433 attempting var/cmdset lookup using [a] 1.096819 very old or unknown APC model, support will be limited What we need to do here is figure out if the "a" result looks good, or if we need to add "5YW" to the compatibility table.

I created a temporary branch to investigate: https://github.com/clepple/nut/compare/apcsmart_logging

Can you please switch to that branch (it's in my own repository rather than the main one, so that it doesn't spin up a bunch of Buildbot jobs), then rebuild and post just the output after "very old or unknown APC model" from the new branch?

[surr]

   1.145806 APC_CMDSET: (86 bytes) => 34 2e 21 24 25 2b 3f 3d 23 7c 2e 01 05 09 0c 0e
   1.145816  14 16 1a 27 29 2b 2d 2f 37 38 39 3c 3e 40 41 42 43 44 45 46 47 4b 4c 4d 4e
   1.145824  4f 50 51 52 53 55 56 57 58 59 5a 5c 5e 61 62 63 65 66 67 6a 6b 6c 6d 6e 6f
   1.145832  70 71 72 73 75 77 78 79 7a 7e 7f 2e 81 3a d0 e2 f0 fd fe ff

[clepple]

It looks like the APC_CMDSET_FMT regex is only expecting two "." characters, and there are three:

1.145806 APC_CMDSET: (86 bytes) => 34 2e 21 24 25 2b 3f 3d 23 7c 2e 01 05 09 0c 0e 1.145816 14 16 1a 27 29 2b 2d 2f 37 38 39 3c 3e 40 41 42 43 44 45 46 47 4b 4c 4d 4e 1.145824 4f 50 51 52 53 55 56 57 58 59 5a 5c 5e 61 62 63 65 66 67 6a 6b 6c 6d 6e 6f 1.145832 70 71 72 73 75 77 78 79 7a 7e 7f 2e 81 3a d0 e2 f0 fd fe ff I don't have any other logs showing what this should look like, but maybe I can find some more info in the apcupsd documentation.=

[clepple]

On 2015-01-31 17:00, Charles Lepple wrote:

It looks like the APC_CMDSET_FMT regex https://github.com/networkupstools/nut/blob/master/drivers/apcsmart.h#L146 is only expecting two "." characters, and there are three:

| 1.145806 APC_CMDSET: (86 bytes) => 342e 21 24 25 2b 3f 3d 23 7c2e 01 05 09 0c 0e 1.145816 14 16 1a 27 29 2b 2d 2f 37 38 39 3c 3e 40 41 42 43 44 45 46 47 4b 4c 4d 4e 1.145824 4f 50 51 52 53 55 56 57 58 59 5a 5c 5e 61 62 63 65 66 67 6a 6b 6c 6d 6e 6f 1.145832 70 71 72 73 75 77 78 79 7a 7e 7f2e 81 3a d0 e2 f0 fd fe ff| I don't have any other logs showing what this should look like, but maybe I can find some more info in the apcupsd documentation.

Sorry for late reply. Yes it seems to be the cause of problem. Current version assumed that command set was returned by 'a' was:

version.alerts.commands

Old driver searched for '.' from the beginning of the string, the new one verified by regex and searched for '.' from the back, obviously failing, as new(?) version is apparently:

version.alerts.commands.something_apc_added

Version was also bumped to 4.

I'll fix it tommorow.

Michal Soltys (msoltyspl)

[msoltyspl]

Pushed small update, it should be working fine now.

Sorry for delay.

[surr]

I get a SEGV now.

Network UPS Tools - APC Smart protocol driver 3.1 (2.7.2-signed-163-g07e2f13)
APC command table version 3.1
   0.000000 debug level is '1'
   0.119102 attempting firmware lookup using [V]
   0.199099 detected firmware version: 5YW
   0.199221 attempting var/cmdset lookup using [a]
   1.139119 parsing out supported cmds/vars
   1.139136 preread_data: ups.model [0x01]
   1.519062 ups.model [0x01] - variable supported
   1.519089 [0x05] unrecognized
   1.519106 [0x09] unrecognized
   1.519123 preread_data: ups.display.language [0x0c]
   1.589085 ups.display.language [0x0c] - variable supported
   1.589114 load.on [0x0e] - command supported
   1.589121 [0x14] unrecognized
   1.589125 [0x16] unrecognized
   1.589130 preread_data: output.current [/]
   1.699069 output.current [/] - variable supported
   1.699096 [7] unrecognized
   1.699101 preread_data: input.quality [9]
   1.769024 input.quality [9] - variable supported
   1.769034 preread_data: battery.packs.bad [<]
   1.859486 battery.packs.bad [<] - variable supported
   1.859510 preread_data: battery.packs [>]
   1.999471 battery.packs [>] - variable supported
   1.999490 shutdown.return [@] - command supported
   1.999497 test.panel.start [A] - command supported
   1.999502 preread_data: battery.voltage [B]
   2.109059 battery.voltage [B] - variable supported
   2.109077 preread_data: ups.temperature [C]
   2.219043 ups.temperature [C] - variable supported
   2.219060 calibrate.start [D] - command supported
   2.219067 calibrate.stop [D] - command supported
   2.219071 preread_data: ups.test.interval [E]
   2.349494 ups.test.interval [E] - variable supported
   2.349522 preread_data: input.frequency [F]
   2.459033 input.frequency [F] - variable supported
   2.459068 preread_data: input.transfer.reason [G]
   2.519031 input.transfer.reason [G] - variable supported
   2.519048 shutdown.stayoff [K] - command supported
   2.519054 preread_data: input.voltage [L]
   2.659036 input.voltage [L] - variable supported
   2.659054 preread_data: input.voltage.maximum [M]
   2.769051 input.voltage.maximum [M] - variable supported
   2.769066 preread_data: input.voltage.minimum [N]
   2.939446 input.voltage.minimum [N] - variable supported
   2.939456 preread_data: output.voltage [O]
   3.049022 output.voltage [O] - variable supported
   3.049038 preread_data: ups.load [P]
   3.219003 ups.load [P] - variable supported
   3.219034 shutdown.return [S] - command supported
   3.219054 test.failure.start [U] - command supported
   3.219075 test.battery.start [W] - command supported
   3.219093 test.battery.stop [W] - command supported
   3.219111 preread_data: ups.test.result [X]
   3.289414 ups.test.result [X] - variable supported
   3.289429 load.off [Z] - command supported
   3.289435 [\] unrecognized
   3.289440 bypass.start [^] - command supported
   3.289446 bypass.stop [^] - command supported
   3.289451 preread_data: ups.id [c]
   3.438983 ups.id [c] - variable supported
   3.438993 preread_data: battery.charge.restart [e]
   3.519020 battery.charge.restart [e] - variable supported
   3.519043 preread_data: battery.charge [f]
   3.639433 battery.charge [f] - variable supported
   3.639457 preread_data: battery.voltage.nominal [g]
   3.719423 battery.voltage.nominal [g] - variable supported
   3.719436 preread_data: battery.runtime [j]
   3.828999 battery.runtime [j] - variable supported
   3.829016 preread_data: battery.alarm.threshold [k]
   3.939408 battery.alarm.threshold [k] - variable supported
   3.939421 preread_data: input.transfer.low [l]
   4.039397 input.transfer.low [l] - variable supported
   4.039407 preread_data: ups.mfr.date [m]
   4.189403 ups.mfr.date [m] - variable supported
   4.189415 preread_data: ups.serial [n]
   4.449398 ups.serial [n] - variable supported
   4.449417 preread_data: output.voltage.nominal [o]
   4.549396 output.voltage.nominal [o] - variable supported
   4.549411 preread_data: ups.delay.shutdown [p]
   4.629411 ups.delay.shutdown [p] - variable supported
   4.629421 preread_data: battery.runtime.low [q]
   4.699394 battery.runtime.low [q] - variable supported
   4.699409 preread_data: ups.delay.start [r]
   4.829814 ups.delay.start [r] - variable supported
   4.829829 preread_data: input.sensitivity [s]
   4.889373 input.sensitivity [s] - variable supported
   4.889398 preread_data: input.transfer.high [u]
   4.989367 input.transfer.high [u] - variable supported
   4.989392 [w] unrecognized
   4.989409 preread_data: battery.date [x]
   5.149380 battery.date [x] - variable supported
   5.149407 preread_data: ups.firmware [b]
   5.289440 ups.firmware [b] - variable combination supported
   5.289465 parsing out caps
   7.659747 output.voltage.nominal [o(W)] - capability supported
   7.659803 input.transfer.high [u(W)] - capability supported
   7.659826 input.transfer.low [l(W)] - capability supported
   7.659847 battery.charge.restart [e(4)] - capability supported
   7.659877 input.sensitivity [s(4)] - capability supported
   7.659887 battery.runtime.low [q(4)] - capability supported
   7.659918 ups.delay.shutdown [p(4)] - capability supported
   7.659946 battery.alarm.threshold [k(4)] - capability supported
   7.659962 ups.delay.start [r(4)] - capability supported
   7.659985 ups.test.interval [E(4)] - capability supported

Program received signal SIGSEGV, Segmentation fault.
0x0000000000404673 in apc_getcaps (qco=0) at apcsmart.c:1007
1007            } else if (vt->flags & APC_PACK)
(gdb) bt
#0  0x0000000000404673 in apc_getcaps (qco=0) at apcsmart.c:1007
#1  0x0000000000404f13 in getbaseinfo () at apcsmart.c:1262
#2  0x0000000000406a6c in upsdrv_initinfo () at apcsmart.c:2067
#3  0x0000000000407d4f in main (argc=0, argv=0x7fffffffeb88) at main.c:633

[surr]

Local variables at the fail location.

(gdb) info locals 
ptr = 0x7fffffffe371 "\360\064\065\065AUTO 60/.160/3 50/.150/3 "
entptr = 0x7fffffffe375 "AUTO 60/.160/3 50/.150/3 "
upsloc = 87 'W'
temp = "#oW53230240200208220uW43253242265276lW43161196184173e4820015253550607590s411Aq4820205071012151820p483020060120240480720960000k4410TLNr483000020060120240480720960E443336168ON OFF\360\064\065\065AUTO 60/.160/3 50/.1"...
cmd = -16 '\360'
loc = 52 '4'
etmp = "OFF", '\000' <repeats 28 times>
endtemp = 0x7fffffffe38e ""
nument = 5
entlen = 5
i = 4
matrix = 0
ret = 206
valid = 0
vt = 0x0
__func__ = "apc_getcaps"

[msoltyspl]

SEGV should be fixed. Check if it works fine when you have the opportunity.

[surr]

It works fine now. Thanks.