summaryrefslogtreecommitdiff
Commit message (Collapse)AuthorAgeFilesLines
* Bump to cxmanage v0.5.1v0.5.1v0.5.xGeorge Kraft2012-10-041-1/+1
|
* Bump pyipmi requirement to 0.5.0George Kraft2012-09-202-2/+2
|
* CXMAN-33: Print an appropriate error if IPMItool is missingGeorge Kraft2012-09-201-5/+12
|
* CXMAN-36, CXMAN-76: Make "cxmanage info" display all versionsGeorge Kraft2012-09-204-9/+32
| | | | This includes hb_spif, socman, a9boot, and u-boot.
* Use subprocess.Popen() instead of subprocess.check_output()v0.5.0George Kraft2012-09-142-3/+5
|
* Remove a debug print for --ipmipathGeorge Kraft2012-09-141-1/+0
|
* bump to version 0.5.0George Kraft2012-09-142-9/+11
| | | | Also bump pyipmi requirement to 0.4.0
* CXMAN-14: Use "file" to verify image typesGeorge Kraft2012-09-141-2/+14
| | | | | We used to do this anyway, but I removed it for windows support. Of course, Cygwin has the "file" tool anyway so we should be ok.
* CXMAN-91: Print results and errors on keyboard interruptGeorge Kraft2012-09-141-14/+26
| | | | | If a command is aborted by keyboard interrupt, we can at least print the results/errors of the nodes that completed.
* Clean up --all-nodes error output slightlyGeorge Kraft2012-09-142-3/+2
|
* cxmanage: Add a runtime check for pyipmi versionGeorge Kraft2012-09-132-2/+20
| | | | | In most cases this isn't really useful -- pip and/or setup.py should take care of it. Oh well.
* Add error tolerance to firmware downloadsGeorge Kraft2012-09-111-7/+20
| | | | | This was done already for uploads, but we'll do it for downloads as well since they also occur during the firmware update process.
* Use time() instead of a counter for firmware transfer deadlinesGeorge Kraft2012-09-111-9/+11
|
* CXMAN-84: Attempt firmware uploads 3 times per partitionGeorge Kraft2012-09-061-17/+22
| | | | | | | | | This should hopefully make firmware updates more resilient during unfavorable conditions where transfers occasionally fail. Also reduced the timeout for a single attempt to 3 minutes. It really should never take longer than about 20 seconds for a single partition anyway.
* CXMAN-85: Timeout after 5 minutes on firmware transfersGeorge Kraft2012-09-041-0/+7
|
* fwupdate/cxpackage: Add new SIMG --version optionGeorge Kraft2012-08-282-4/+11
|
* simg: Add new SIMG version fieldGeorge Kraft2012-08-283-18/+45
| | | | and bump hdrfmt to 2
* cxmanage: Rename simg "version" to "priority"George Kraft2012-08-288-65/+69
| | | | | This includes renaming the fwupdate/cxpackage "--version" option to "--priority".
* target: Abort ipinfo/macaddr commands if they return an errorGeorge Kraft2012-08-241-5/+9
| | | | | Also cleaned up some other error messages -- they should hopefully be more meaningful now.
* cxmanage: Allow --retry option to work with any commandGeorge Kraft2012-08-242-89/+118
|
* cxmanage: Check ipmitool version at startGeorge Kraft2012-08-231-1/+25
|
* Make "info dump" command run in parallelGeorge Kraft2012-08-233-118/+171
|
* controller: Make output more consistent between commandsGeorge Kraft2012-08-231-30/+15
| | | | | Always print a list of discovered IP addresses, and always print the current action just above the indicator.
* cxmanage: Don't allow infinite-depth hostfile recursionGeorge Kraft2012-08-222-51/+70
| | | | | | | | | Hostfiles can contain any entry that the hostname argument could contain... including entries for other hostfiles. So don't allow a single hostfile to be parsed more than once, in case of circular dependencies. Also allow "file=" as well as "hostfile=" for hostfile entries.
* When using --all-nodes, obtain IP addresses in parallel.George Kraft2012-08-223-39/+73
| | | | | | | | Getting IP addresses from the fabric can be slow since it has to wait for a TFTP transfer. It's generally not noticable with just a single host, but for multiple fabrics, it can take a while. Getting addresses in parallel should speed things up a bit.
* bump to version 0.4.0v0.4.0George Kraft2012-08-211-2/+2
|
* Add Calxeda license to cid / reg ID filesGeorge Kraft2012-08-213-2/+71
|
* cxmanage: Don't allow --threads to be less than 1George Kraft2012-08-211-0/+2
|
* Replace some lingering instances of "slot" with "partition"George Kraft2012-08-203-17/+17
|
* cxmanage: Print an appropriate error message when --all-nodes failsGeorge Kraft2012-08-201-2/+8
|
* infodump: Disable the ipmitool session info commandGeorge Kraft2012-08-201-2/+3
| | | | Seems to cause an MPU fault on real hardware
* target: Use IPMITOOL_PATH env variable for arbitrary "ipmitool" cmdGeorge Kraft2012-08-201-3/+9
| | | | | Also, always print ipmitool output even when it returns exit status 1.
* cxmanage: Add --ipmipath optionGeorge Kraft2012-08-171-0/+9
| | | | | | | | For specifying a path to ipmitool. Can be a directory or a direct path. For example, both of these will work: cxmanage --ipmipath /usr/bin fwinfo 192.168.100.250 cxmanage --ipmipath /usr/bin/ipmitool fwinfo 192.168.100.250
* cxmanage: Add --retry COUNT optionGeorge Kraft2012-08-172-13/+31
| | | | | | | | | | | This option will cause cxmanage to retry the command a fixed number of times on failed hosts, instead of prompting. This will probably be useful for very large clusters where failures are more likely to occur. For example, to try the config reset command 3 times (2 retries): cxmanage --retry 2 config reset
* target: Some improvements to partition selectionGeorge Kraft2012-08-171-34/+20
| | | | | | Mainly, when selecting an "INACTIVE" partition, take into account the valid flag. Also use sorting instead of branching if statements.
* cxmanage: Add "fwinfo" commandGeorge Kraft2012-08-172-1/+33
| | | | | For convenience. This is pretty much just a shortcut to the ipmitool cxoem fw info command.
* controller: For most commands, prompt to retry on failed hostsGeorge Kraft2012-08-161-62/+54
| | | | | This only applies to commands that change the host in some way -- fwupdate, config reset, power on, etc.
* target: Use fwinfo "partition" response field instead of "slot"George Kraft2012-08-163-16/+17
| | | | Also update controller_test and target_test with these changes.
* cxmanage: Make sensor_name argument be a filterGeorge Kraft2012-08-163-12/+23
| | | | | So you can say, for example, "cxmanage sensor power" and it'll list all sensors with "power" in the name.
* cxmanage: Allow "sensor" command with no sensor_name argumentGeorge Kraft2012-08-161-2/+3
| | | | | In this case it's exactly like "sensor list" -- it'll just display all of the sensors.
* target: Use more appropriate error messages for IPMI errorsGeorge Kraft2012-08-151-20/+27
|
* cxmanage: Rename --tftp argument to --external-tftpGeorge Kraft2012-08-151-13/+14
|
* cxmanage: Rename fwupdate --slot argument to --partitionGeorge Kraft2012-08-153-116/+120
| | | | Also updated variable names and such to be consistent.
* controller: Only add skip_crc32 to package manifests when it's trueGeorge Kraft2012-08-101-1/+3
| | | | Also print the skip_crc32 flag along with other image info.
* simg: Use 4KB alignment for CDB and boot log imagesGeorge Kraft2012-08-102-14/+30
|
* cxmanage: Add option to read hosts from a hostfileGeorge Kraft2012-08-081-11/+26
| | | | | | | | | | | | This is specified by using a "hostfile=" entry in the hostname argument. For example, to run the info command using a file called "example_file": cxmanage info hostfile=example_file The hostfile itself can just be a simple list of entries, separated by spaces, newlines, or commas. The # character can be used for comments.
* Various cleanup in the cxmanage scriptGeorge Kraft2012-08-081-40/+42
|
* cxmanage_test: Clean up the controller and target testsGeorge Kraft2012-08-072-130/+63
|
* target: Change method names so they make more senseGeorge Kraft2012-08-074-91/+92
|
* cxmanage_test: Add tests for info_basic and info_ubootenvGeorge Kraft2012-08-062-0/+70
|
View Lorry Controller Status