2
0
mirror of https://github.com/xcat2/xcat-core.git synced 2025-05-29 09:13:08 +00:00

72 Commits

Author SHA1 Message Date
Bin Xu
eea661e405 merge from master to 2.13 branch for 2.13.9 release. (1) (#4525)
* fix issueNode range not specified, see man page for syntax. return with no output when site.master is not set #4299

* Fix issue 4246, record rflash process in log file

* dodiscovery: better disksize - ordered by major,minor and no `\n` (#4219)

* better disksize: ordered by major,minor and no `\n`

* nodediscover:`disksize` sent correctly: no need to `split`,`join`

* general sort with explicit key columns, fix search pattern

* dodiscovery: Fix bugs

* dodiscovery: Fix ShellCheck SC2007,SC2004

* dodiscovery: simpler kcmdline parsing, disksize as IEC binary prefix

* Add test case - switch_to_dns_forward_mode

* Modify timeout of login by curl command for OpenBMC

* Add makegocons command

This patch enable goconserver service and manage the node sessions
with a new `makegocons` command.

Implement: #4215

* Refine the Cumulus Linux Section of the Doc (#4249)

* No content in the switches subdirectory in docs, remove

* Change the Network topic to Networking

* Refine and reformat the Cumulus Linux documentation

* Modify the script to build xCAT-genesis-base package (#4292)

* Verify hash ID

* change log messages

* Timeout the ntpd process if ntpd service is not running on ntpserver (#4321)

* enhance nodediscovery process: don't write "NOIP" if the node can not be resloved to an IP (#3995)

* enhance nodediscovery process, if only 1 mac and have \*NOIP\* append, don't write mac table and don't generate dhcp lease entry

* Enhance PR 3995: enhance nodediscovery process: don't write "NOIP" if the node can not be resloved to an IP

* modified depending on comments

* natural_sort_cmp: recursion to iterative implementation (#4314)

* natural_sort_cmp: recursion to iterative implementation

* natural_sort_cmp: correct prototype with parameters, as per `man perlfunc`

* Task 3339, rspconfig ip/netmask/gateway/vlan support

*   * Minor enhance on xcatperftest to put all logs into one file
  * Fix a bug in simulatorctl.sh, and it cause the script cannot be found

* modified depending on comments

* QA list for makehosts (#4284)

* QA list for makehosts

* Fix some bus in makegocons

* Revise test case - switch_to_dns_forward_mode

* Enhancements after the review

* Use IO::Socket to check BMC console port

* Revise test case - switch_to_dns_forward_mode

* modified depending on comments

* Add messages to inform the user of the general action started via xCAT for flashing firmware

* Fix bug, anything in the functional array is the one that's really active, priority will not be 0 if there is pending firmware

* Fix 4338, remove all of  in child process

* fix issue #4354 :The XCATTEST_CN in xcattest can not detect HCP as config file (#4355)

* Fix #4330, close the socket

* Add image name to /opt/xcat/xcatinfo on compute node (#4359)

* Integrate congo console from goconserver with rcons

Enhance the original implement of rcons to support goconserver.
`rcons` will help the user choose from one of the console backend
service based on the console server status of management node.

Implement #4216

* Fix 4363 - discovery broken (#4364)

* do not set exit_code to 1 when the clock is not synced before the timeout (#4362)

* Fix check for MTM values with spaces

* modified depending on comments

* rspconfig admin_password for OpenBMC

* Fix merging in xdcpmerge.sh (#4328)

* Fixes in xdcpmerge.sh

Two fixes:
1. The grep pattern when finding duplicate usernames is missing ":" at the end. So, for example user "test" would also match "test2, etc.". Adding the ":" delimiter fixes the issue.
2. Another issue happens when the file to be merged is a superset of the files on the nodes. For example, if a new user is added and entire passwd file (that is otherwise identical) is sent to be merged. In this case, the $filebackup.nodups file, i.e. the original file with duplicates removed, becomes empty and the condition "if [ -s "$filebackup.nodups" ]" does not execute. Then the merged file ends up being original file with the merge file fully appended, clearly not what was intended.

This is solved by changing the condition to check for file existence "-a" rather then for size. Additionally, I also turn the logic around so that the duplicates are removed from the merge file and then added to the original file. I think this makes logic a bit cleaner and also ensures that existing entries are not reordered or changed in any way.

* Streamlining previous commit

Adjustment to previous commit, streamlining and simplifying logic. Once $mergefile.nodups is created, just concatenate it the original file.

* Update to xdcpmerge

No need to copy $filebackup to $curfile, they are the same.

* Modify for debug conveniently

* add new cases and delete outdated test cases

* Modify genesis build script for centos x86_64

* Add space between at and the time

* Add a print out of the firmware levels for the various UT cases

* Add unit test cases for rspconfig

* rspconfig fix for set hostname

* Enhance the testing case for rspconfig setting hostname

* modified depending on comments

* modified depending on comments

* Add %pretrans script in <lua>. Handle directory to symlink change properly. See comment #3 of https://bugs.launchpad.net/rpm/+bug/633636

* Make xCAT-genesis-base confliects with early version of xCAT-genesis-scripts

* rspconfig dump to allow admins capture logs

* Adding comment

* Improve the error message when BMC does not return a dump ID

* Improve some messages and add timestamp for downloaded dump file

* Improve the message to help Admin figure out where the file is missing

* Leave a log file there when xCAT upgrade in case to debug issue while upgrading (#4389)

* Listen on 0.0.0.0 instead of the hostname

This patch modify the configuration of `makegocons` and `rcons`
for goconserver.

`cat /etc/goconserver/server.conf`
```
global:
  host: 0.0.0.0
  ssl_key_file: /etc/xcat/cert/server-key.pem
  ssl_cert_file: /etc/xcat/cert/server-cert.pem
  ssl_ca_cert_file: /etc/xcat/cert/ca.pem
  logfile: /var/log/goconserver/server.log
api:
  port: 12429
console:
  port: 12430
```

* Support hostname=* for openbmc

* Relay action and snmp configuration support for Coral PDU

* ddns.pm: specify the "directory" option for DNS slaves too (cf. bug #4392)

* Fix issue 4361, modify some sendmsg to message

* 1. add "makeconserver -d" to "rmdef -C", 2. add "makeconserver -C|--cleanup" to remove entries for removed nodes

* When there is a problem with the login, do not hide the message on debug mode. BMCReady does not make sense if the admin does not know how to find that state

* Change function from login_logout_request to login_request, not doing any logout here

* Check that RC is 200 to prevent unknown issues, handle the response generically

* OpenBMC rspconfig dump timeout fixes

* Fix issue 4408, modify error for rspconfig dump

* Clear all BMC Dump logs when BMC firmware flash

* modified depending on comments

* More modifications for pr 4386, to deal with the conflicts

* To handle one case which have 2 implementations, which one is for specific platform, on is for all platforms

* return when current status is RSPCONFIG_DUMP_DOWNLOAD_REQUEST

* fix issue 4417, delete 'clear next_status'

* fix issue 4353: rspconfig needs to support multiple IPs on the BMC and ignore ZeroConfigIPs

* Wait 15 seconds after OpenBMC interface with vlan tag to be activated

* Fix issue #4397: rspconfig <> hostname=xxx show error message when there is multiple network in bmc

* Some sentence modify for makeconservercf -C|--cleanup

* OpenBMC rspconfig dump better dump file name formatting

* Removed the --check and --ipsource option with PR 4258, update the man page

* Improve the message on the HTTP response

* modified depending on comments

* Only handle 404 and 504 in the login request code, defer the rest to deal_with_response()

* rflash stream support

* 1. configure ip/netmask/gateway only on the NIC whose IP match node BMC attribute, 2. add some information for LinkLocal address

* Fix confignetwork bond nic_type detection with multiple bonds

* Modified configonie --ntp command (#4436)

* Add man page for makegocons

This is the guide about how to make goconserver as
a replacement for conserver to help slove the issues reported
for conserver, like: #4043, #3543. For openbmc, the solution of
goconserver is much light-weighted than the conserver which could
help save the system resource. In addition, sshpass is not needed
for openbmc with goconserver.

Implement: #4337

* Add another key for node_info in order not to after the content of $node_info{$node}{bmc}

* enhance rflash stream

* makedhcp does not work well when all service nodes not running dhcp but disjointdhcps=1 (#4426) (#4440)

- if all service nodes not running dhcp, to treat it as disjointdhcps=0
- nodeset will send request to MN by default even if disjointdhcps=1
- Move out of the dhcp service checking from opts pre-check, and do it just before real makedhcp handling.

* rspconfig configure bmc vlan will hung because of PR 4383

* OpenBMC rspconfig dump enhancements

* Changes due to review comments

* Print debug message before login attempt

* Add warning when xCAT throttles SSL connections

* Display first [openbmc_debug] when entering openbmc.pm

* modified error msg

* Make sure credential files have a trailing newline (#4442)

* modified depending on comments

* Fix the typo in the man page of makegocons

* Update the print out based on the review comment, should not use  since the regular expression is removed

* Modify the nodeset disjoint test case accordingly for #4426

* Use short hostname in rcons for goconserver

As the certificate of xcat is signed with short hostname, this
commit force to use the short hostname in  the environment variable for
`congo console`.

* Fix issue 3497, make sense for reventlog msg

* Give summary after flash active when no debugmode

* Fix the issue that the IP configuration will fail if bmc attribute is a hostname

* enhance genimage for sles12sp2 (#4450)

* Add dhcp-client-identifier to lease block (#4429)

Machines that use Infiniband for PXE booting need to have the
dhcp-client-identifier set in the lease block.
Without it, they will not get the lease from the server.

* Support multiple bonds on bring-up

* modified depending on comments

* fix the check for rc to 1 on error cases

* modify response for bmcdiscover when error

* Ignore syslog error in monitorctrl when setNodeStatusAttributes (#4459)

* fix issue https://github.com/xcat2/xcat-core/issues/4411 (#4462)

* fix issue Compute nodes fail to get provisioned #4411: covert imgsrv and xcatmaster to their ip addresses in case the hostname cannot be resolved inside initrd for diskless

* More strict check to tell if it is a chroot env to avoid modify DB (#4463)

when genimage for SN image (#4365)

* issues for install license file on accton switches (#4460)

* Add test cases for rflash regular usage against openbmc

* modify depending on xuwei's comment

* add 2 more cases for option d

* enhance rflash upload message

* Do not display message for clearing dumps when only PNOR

* Display hostname even if multiple IP addresses

* modified depending on comments

* polished message

* Modify the default consoleondemand based on the global setting

This commit fix the bug that consoleondemand works incorrectly.

* modify depending on comments

* modified depending on comments

* enhance rflash error messages

* Modify documenation for servicenode attributes

* build rst file from Schema.pm by db2man

* change status back to starts

* modified depending on comments

* rm openbmcevents

* Usage and man page update for rspconfig dump

* let rflash error message flexible

* Add support for the "file -> (noderange) file" syntax in synclist with ServiceNodes (#4445)

* Add support for "file -  (noderange) file" in synclist when using
hierarchical mode. Fixes #4425

  This patch ensures that:
  1. the synclist is correctly parsed when running on a Service Node
  2. all files are synchronized to SNs in hierarchical mode

* Better test condition for #4425, addresses issue in
https://github.com/xcat2/xcat-core/pull/4445#issuecomment-349472901

* Fix issue 4477, if has node-<mac> will not create node-<mtms> for the same node

* fix issue updatenode -f loses directories when copying files to SN #4456 (#4494)

* comment from ErTao

* Crude attempt at including external configuration files in named.conf

* Fixes after the review

* Fix issue 4490, record any error when rflash active process

* add -d usage and manpage

* updatenode -F not work in hierachy env as the user name is FQDN of MN (#4484)

* updatenode -F not work in hierachy env (#4455)
 - add trace when -V is enabled
 - get the DSH_FROM_USERID from updatenode client

*  - when 'updatenode -F' need to push SN first, using root as non-root does not have permission write to 'SNsyncfiledir'
 - move the set DSH_FROM_USERID code out of the loop, and also cover remote client case.

* fix issue for command rspconfig hostname=*

* enhance rflash

* Adjust the server used for kernel/initrd and imgurl for petitboot (#4416)
 - URL for kernel/initrd, get the value from below value tftpserver -> xcatmaster -> myipfn
 - URL for image, get the value from below value nfsserver -> tftpserver -> xcatmaster -> myipfn

* NODE attribute didn't populate in /opt/xcat/xcatinfo after reboot (#4428)

* NODE attribute didn't populate in /opt/xcat/xcatinfo after reboot

* Get NODE from mypostscripts

* Improve the output message for reventlog, use a global variable to set PolicyFile Path

* If debug_msg is not provided, use an empty string

* Check for LinkLocal as well as 169.254 IP address

* Fix issue 4507, add parameter check for rspconfig admin_passwd

* record more information when rflash upload error

* Fix the error when using array ref in updatenode with old version perl, it is introduced by PR#4484 (#4518)

* Do not restart conserver if goconserver was started

If goconserver was enabled, do not start conserver when restart
xcatd on service node.

* remove the /etc/localtime before copy timezone file

* Use CONGO_CLIENT_TYPE to tell goconserver the source of client (#4501)

goconserver could send back message based on the client type
this commit set CONGO_CLIENT_TYPE to xcat to make the message
from goconserver more friendly.

* add rflash -d doc

* only ignore 169.254.x.x for OpenBMC

* Fix issue 4513, print out better error msg for reventlog -s

* Modify or add openbmc test cases or bundle

* add test cases for updatenode -f/F in hierarchy environment, covers issues #4456,#4455 and PR #4425 (#4500)
2017-12-14 05:03:34 -06:00
cxhong
f295689c22 Merge pull request #4005 from mattaezell/ztp_improvements
Improvements to Cumulus ZTP
2017-10-02 13:39:29 -04:00
Matt Ezell
a9b23dfa1d Improvements to Cumulus ZTP
- Use the full path to cl-license
- Allow configinterface to download files
- Run enablesnmp in mypostscript so we have the correct community string
2017-09-25 16:04:09 -04:00
immarvin
d8d40f184c add cumulus upgrade doc 2017-09-25 09:38:37 -04:00
chenglch
bc6ddccbd0 Modify the document and nics schema based on review comments. 2017-07-21 16:19:24 +08:00
chenglch
e8df039d42 Make getadapter work during discovery process
This patch redesign the original getadapter plugins. Currently,
getadapter do not work as a common command but only a chain command
works during discovery.
2017-07-20 16:03:52 +08:00
yangsong
2d497a0d8c documentation for ZTP based ONIE switch setup (#3362) 2017-06-29 11:36:23 +08:00
Mark Gurevich
7e2458d7de Spelling fixes for docs/sources/advanced rst files 2017-05-22 16:39:16 -04:00
cxhong
1c51f5dcf4 Supports snmpv3 configuration based on the attributes defined in the switches table (#2956)
* Supports snmpv3 configuration based on the switches table for onie switch

* fix typo
2017-05-05 03:37:56 -05:00
Victor Hu
01a5551ea7 Update manage.rst
Clarify the option to clear the previous configuration for ONIE enabled switches
2017-02-24 13:15:42 -05:00
cxhong
fc95d2e958 Update install.rst 2017-02-23 11:46:07 -05:00
yangsong
0b94414ec1 Merge pull request #2458 from whowutwut/onie
Refactor the ONIE documentation
2017-02-13 20:23:47 -06:00
Victor Hu
8c5195bb97 Add some explanation about how the cumulus linux file gets sent to the switch for installation 2017-02-13 15:11:56 -05:00
Victor Hu
9320fd8a28 Refine some of the instructions for the cumulus switch 2017-02-09 09:49:56 -05:00
Victor Hu
8646369568 add information about configinterface postscript 2017-02-06 11:16:14 -05:00
Victor Hu
af0bd14071 minor Formatting changes for the cumulus section 2017-02-06 10:07:42 -05:00
Victor Hu
df7b859fc4 Refactor the onie documentation 2017-02-03 15:52:42 -05:00
Victor Hu
808a2ed5e9 Rename edgecore_switches to onie_switches 2017-02-03 09:41:13 -05:00
Mark Gurevich
d076feb216 Merge pull request #2454 from whowutwut/security0127
Looks good to me.
2017-01-31 16:49:15 -05:00
Victor Hu
19f6e8d242 fix random errors in documentation build 2017-01-31 16:38:55 -05:00
Victor Hu
1a400c31ab Update mlnxofed_ib_install_v2_diskful.rst 2017-01-31 16:01:55 -05:00
Victor Hu
d1088318bc Update mlnxofed_ib_install_v2_diskless.rst 2017-01-31 16:00:36 -05:00
Casandra Qiu
f843dcad3c Change switchtype from cumulus to onie 2017-01-10 15:38:14 -05:00
immarvin
5c9a3624d6 refine the doc on cumulus switch 2017-01-03 21:44:15 -05:00
immarvin
2e1e273267 refine cumulus support doc 2016-12-23 01:33:05 -05:00
immarvin
4f771e3dc3 refine the cumulus support doc 2016-12-23 01:33:05 -05:00
cxhong
1cd4b6446c Check in init version of cumulus support[Do Not Merge] (#2020)
* Check in init version of cumulus support

* Add cumulus installation doc

* modify victor's comments.
2016-11-08 10:21:56 +08:00
Mark Gurevich
a50f3710f7 Doc spelling and gramma fixes 2016-10-25 15:31:20 -04:00
Weihua Hu
0c951db82b Merge pull request #1945 from whowutwut/mellanox_documentation_improvements
Mellanox IB documentation clarification for OFED Installation Step
2016-10-10 14:45:06 +08:00
Victor Hu
ad4d60627a Move the kernel mismatch issue back to the diskless and diskful sections 2016-10-07 15:56:12 -04:00
Victor Hu
87b8b1ea52 Reformat some of the information in the known issues 2016-10-07 14:05:34 -04:00
Victor Hu
e10b5afad3 Refactor the mellanox installation instructions to make it easier to follow 2016-10-07 13:38:25 -04:00
Victor Hu
ffba03e3bc Remove the documentation for mellanox ofed installer version 1.
We have version control in the documentation, so there's no need to
keep the deprecated documentation in the source
2016-10-05 16:04:52 -04:00
Matt Ezell
1e11b6e80f Add support for a default MTU on a network
Add a MTU column to the networks table.  If set, this will configure the MTU in
DHCP so that nodes will have the correct MTU during provisioning.  This is
important for large systems to enable jumbo frames to reduce CPU usage when
copying diskless images over the network (before the confignics postscript can
run to set the desired MTU).
2016-10-01 16:01:28 -04:00
Casandra Qiu
019c0c6aa8 Support Predefine switch has same hostname as discovery 2016-09-22 21:35:04 -04:00
Casandra Qiu
3b1785802f modify from Mark's comment 2016-09-19 12:46:16 -04:00
Casandra Qiu
e7726d67da Modify comments from Mark 2016-09-15 21:03:24 -04:00
Casandra Qiu
09b578137a modify the doc with comment 2016-09-13 16:29:50 -04:00
Casandra Qiu
b90ca84237 Modify switchdiscover usage, man page and update readdoc 2016-09-12 22:42:41 -04:00
Mark Gurevich
ef3bf0c437 Fix doc build errors and missing links 2016-05-24 16:16:34 -04:00
hu-weihua
c950ca7507 modify spelling error 2016-04-18 02:24:50 -04:00
hu-weihua
b7beeabcf8 modify format error 2016-04-18 02:15:08 -04:00
hu-weihua
49a85041fe replace infiniband to InfiniBand 2016-04-15 04:07:25 -04:00
hu-weihua
9ffa8e314d modify grammar error dependig on gongjie's comments 2016-04-15 03:44:38 -04:00
hu-weihua
b2a8ddecf0 update example 2016-04-14 03:25:54 -04:00
hu-weihua
642059b3c2 modify more content for bug fixing 2016-04-12 03:18:24 -04:00
Casandra Qiu
4488c24a30 Add snmp_scan method for switchdiscover command 2016-03-24 21:22:58 -04:00
Victor Hu
7ae8361209 Issue #766, Fix a typo in the switch configuration command 2016-03-11 10:58:07 -05:00
Victor Hu
d75ecae303 Fix errors/warnings in the documentation during the build 2016-03-03 13:36:44 -05:00
chenglch
c6e663d1fd Use fifo pipe to enhance getadapter
Currently the getadapter implementation store the nics information
in multiple files. This implementaion will store the records in the
database and add the following changes:

1. Add taskstate table to record the getadapter request.
2. Add route_request hook in xcatd to avoid of too much process.
3. Add adapterinfo column to keep the adapter information.
4. Use fifo pipe to send/recv the adapter message and wake up the waiting process.
5. Use alarm signal to handle the timeout event.
2016-01-25 01:53:54 -05:00