ISC Open Source Projects issueshttps://gitlab.isc.org/groups/isc-projects/-/issues2019-04-06T19:23:38Zhttps://gitlab.isc.org/isc-projects/bind9/-/issues/973Deadlock in RPZ update code2019-04-06T19:23:38ZWitold KrecickiDeadlock in RPZ update codeIn dns_rpz_update_from_db we call setup_update which creates the db iterator and calls dns_dbiterator_first. This unpauses the iterator and might cause db->tree_lock to be acquired. We then do isc_task_send(...) on an event to do quantum...In dns_rpz_update_from_db we call setup_update which creates the db iterator and calls dns_dbiterator_first. This unpauses the iterator and might cause db->tree_lock to be acquired. We then do isc_task_send(...) on an event to do quantum_update, which (correctly) after each iteration calls dns_dbiterator_pause, and re-isc_task_sends itself.
That's an obvious bug, as we're holding a lock over an async task send - if a task requesting write (e.g. prune_tree) is scheduled on the same workers queue as update_quantum but before it, it will wait for the write lock indefinitely, resulting in a deadlock.
To fix it we have to pause dbiterator in setup_update.BIND 9.14.1Witold KrecickiWitold Krecickihttps://gitlab.isc.org/isc-projects/kea/-/issues/559[ISC-support #17133] Add Dockerfile and publish to dockerhub2023-09-28T08:50:16ZGhost User[ISC-support #17133] Add Dockerfile and publish to dockerhub---
name: Add Dockerfile and publish to dockerhub
about: Ease of install
---
@SupportRT
RT [#17133](https://support.isc.org/Ticket/Display.html?id=17133)
**UPDATE**: There are several Dockerfiles available:
- Shared in #1109: https:...---
name: Add Dockerfile and publish to dockerhub
about: Ease of install
---
@SupportRT
RT [#17133](https://support.isc.org/Ticket/Display.html?id=17133)
**UPDATE**: There are several Dockerfiles available:
- Shared in #1109: https://github.com/sjmiller609/docker-images
- Shared in #1233 by @JonasAlfredsson: https://github.com/JonasAlfredsson/docker-kea
- Kea images used by Stork: https://gitlab.isc.org/isc-projects/stork/-/tree/master/dockerkea2.5.2https://gitlab.isc.org/isc-projects/bind9/-/issues/974dlz dlopen driver always uses RELATIVERDATA due to faulty flags handling2021-10-04T19:11:05ZCathy Almonddlz dlopen driver always uses RELATIVERDATA due to faulty flags handlingAs noted in [Support ticket # 14218](https://support.isc.org/Ticket/Display.html?id=14218#txn-474621) There is a bug in flags handling in dlz dlopen driver, causing it to always use RELATIVERDATA.
The fix is unfortunately not trivial. A...As noted in [Support ticket # 14218](https://support.isc.org/Ticket/Display.html?id=14218#txn-474621) There is a bug in flags handling in dlz dlopen driver, causing it to always use RELATIVERDATA.
The fix is unfortunately not trivial. A much easier workaround is making sure that a DLZ module always adds absolute names.
9.11https://gitlab.isc.org/isc-projects/kea/-/issues/561Make kea to build on last macOS / Xcode2019-05-23T09:40:16ZFrancis DupontMake kea to build on last macOS / XcodeI updated macOS, Xcode, all brewed dependencies, libyang/sysrepo. I am trying to make Kea with premium to build with all options on. Unfortunately the last clang (1001.0.46.3) does not accept some features.
This ticket is to fix all iss...I updated macOS, Xcode, all brewed dependencies, libyang/sysrepo. I am trying to make Kea with premium to build with all options on. Unfortunately the last clang (1001.0.46.3) does not accept some features.
This ticket is to fix all issues which block build and make check on macOS (the idea is they should block on some other systems too).Kea1.6Francis DupontFrancis Duponthttps://gitlab.isc.org/isc-projects/bind9/-/issues/975pkcs11 slot number relation to dnssec-keyfromlabel URI2021-10-04T19:12:00ZPetr Menšíkpkcs11 slot number relation to dnssec-keyfromlabel URI### Description
I were unable to configure automated testing of system test pkcs11. Both pkcs11 and pkcs11ssl failed in my testing. dnssec-keylabel with --enable-native-pkcs11 accepts pkcs11 URI, which is great. However, support pkcs11 ...### Description
I were unable to configure automated testing of system test pkcs11. Both pkcs11 and pkcs11ssl failed in my testing. dnssec-keylabel with --enable-native-pkcs11 accepts pkcs11 URI, which is great. However, support pkcs11 tool pkcs11-keygen accepts only slot number. When I was scripting our pkcs11 build with [custom patch](https://src.fedoraproject.org/rpms/bind/c/25e332108e68a819adc334c9c3261a9d8188bf71?branch=master) and [helper softhsm script](https://src.fedoraproject.org/rpms/bind/blob/master/f/setup-named-softhsm.sh), I failed to find correct parameters.
I think my issue is there is no slot 0 initialized by the script. That can be overriden by SLOT environment variable. However, if I have one initialized token and one unintialized, dnssec-keyfromlabel does not know which one to use. I did not find way to specify token in pkcs11 URI by slot number.
Better way would be to support pkcs11 uri to select token in pkcs11 tools too.
```
$ p11tool --list-all
pkcs11:model=PKCS%2315%20emulated;manufacturer=piv_II;serial=*;token=Petr%20Mensik
pkcs11:model=SoftHSM%20v2;manufacturer=SoftHSM%20project;serial=*;token=DNS
pkcs11:model=SoftHSM%20v2;manufacturer=SoftHSM%20project;serial=*;token=test
```
Is there way to use ``$SLOT`` variable that I did not find? It fails to generate private key, because it cannot find correct token.
### Request
- Support PKCS11 URI in all tools if possible
- Document way to specify slot number to dnssec-keygen in compatible way with pkcs11 tools, if there is any
- Provide clear token not found return code or message in pkcs11-* tools
- Provide login failed message in pkcs11-* tools on bad HSM pin
- Ignore uninitialized tokens altogether
### Links / references
- Fedora guidelines demands PKCS11 URI support for any tool working with tokens. [Fedora Packaging policy of PKCS#11](https://docs.fedoraproject.org/en-US/packaging-guidelines/Pkcs11Support/). Would like to provide support for p11-kit integration when I find enough time for it.
- Found no way to supply setup parameters of [pkcs11 setup](https://gitlab.isc.org/isc-projects/bind9/blob/master/bin/tests/system/pkcs11/setup.sh)
- [Softhsm setup script](https://src.fedoraproject.org/rpms/bind/blob/master/f/setup-named-softhsm.sh)https://gitlab.isc.org/isc-projects/kea/-/issues/562Rate-limit for DHCPv62022-06-21T15:00:17ZTomek MrugalskiRate-limit for DHCPv6A customer is requesting Kea ability to rate-limit the leasing IPv6 addresses:
> Furthermore, we are seeing that DHCPv6 allows a single device
> to request multiple IP addresses if it can randomly change the
> DUID. Is there some mechan...A customer is requesting Kea ability to rate-limit the leasing IPv6 addresses:
> Furthermore, we are seeing that DHCPv6 allows a single device
> to request multiple IP addresses if it can randomly change the
> DUID. Is there some mechanism to rate-limit the leasing of IPs
> given to the same device?
Although this particular request is v6 specific, the same concept can be easily extended to v4.kea2.1.6https://gitlab.isc.org/isc-projects/bind9/-/issues/976dns/ecs.h missing ISC_LANG_ENDDECLS2019-04-09T01:59:31ZMark Andrewsdns/ecs.h missing ISC_LANG_ENDDECLSMark AndrewsMark Andrewshttps://gitlab.isc.org/isc-projects/kea/-/issues/563Remove obsolete dependency in dbaccess_parser.cc2019-07-22T10:53:04ZFrancis DupontRemove obsolete dependency in dbaccess_parser.ccsrc/lib/database/dbaccess_parser.cc has an obsolete (i.e. still compiles when it is removed and no obvious use in the code) dependency on dhcpsrv/parsers/dhcp_parsers.h. I propose to remove it as the database library is built before the ...src/lib/database/dbaccess_parser.cc has an obsolete (i.e. still compiles when it is removed and no obvious use in the code) dependency on dhcpsrv/parsers/dhcp_parsers.h. I propose to remove it as the database library is built before the dhcpsrv one.
BTW a comment in db_log.h needs to be updated too according to a grep dhcpsrv in the directory.Kea1.6-beta2Francis DupontFrancis Duponthttps://gitlab.isc.org/isc-projects/dhcp/-/issues/6DHCPv6 lease length logging2021-03-02T07:27:17ZGhost UserDHCPv6 lease length logging---
name: DHCPv6 - on commit {} - preferred lifetime / valid lifetime incorrect values available
---
**Describe the bug**
When logging preferred and valid lifetime using on commit {} client requested values are recorded instead of serve...---
name: DHCPv6 - on commit {} - preferred lifetime / valid lifetime incorrect values available
---
**Describe the bug**
When logging preferred and valid lifetime using on commit {} client requested values are recorded instead of server provided values.
**To Reproduce**
Use a MacOS DHCPv6 client.
Add this bit to the dhcp config:
```
on commit {
if exists dhcp6.ia-na {
log(debug,
concat( "PREFERREDLIFETIME: ",binary-to-ascii(10, 32, "", substring(option dhcp6.ia-na,32,4)),",",
"VALIDLIFETIME: ",binary-to-ascii(10, 32, "", substring(option dhcp6.ia-na,36,4))
)
);
}
}
```
which prints a log line like this:
`Sep 15 18:53:49 dhcp-server-1 dhcpd: PREFERREDLIFETIME: 0,VALIDLIFETIME: 0`
**Expected behavior**
Should print a log line with the server provided preferred and valid lifetimes (ie: the values being sent back to the client. In my test case was 375 and 600)
**Environment:**
- ISC DHCP version: 4.4.1
- OS: Linux (custom)
- Which features were compiled in
```
./configure --prefix=/usr \
--sysconfdir=/etc \
--enable-secs-byteorder \
--localstatedir=/var/state/dhcp \
--with-srv-lease-file=/var/state/dhcp/dhcpd.leases \
--with-srv6-lease-file=/var/state/dhcp/dhcpd6.leases \
--with-srv-pid-file=/var/run/dhcpd.pid \
--with-srv6-pid-file=/var/run/dhcpd6.pid;
```
**Additional Information**
when committing a lease to an Apple Mac mini (el Capitan) which generates a conventional log line like this:
```
Sep 15 18:53:48 dhcp-server-1 dhcpd: Relay-forward message from 2001:DB8:2e50:e8::1 port 547, link address 2001:DB8:2e50:e8::1, peer address fe80::225:4bff:fea0:6fe8
Sep 15 18:53:48 dhcp-server-1 dhcpd: Advertise NA: address 2001:DB8:2e50:e8:7fff:ffff:ffff:fffe to client with duid 00:01:00:01:20:e5:6e:2d:00:25:4b:a0:6f:e8 iaid = 0 valid for 600 seconds
Sep 15 18:53:48 dhcp-server-1 dhcpd: Sending Relay-reply to 2001:DB8:2e50:e8::1 port 547
Sep 15 18:53:49 dhcp-server-1 dhcpd: Relay-forward message from 2001:DB8:2e50:e8::1 port 547, link address 2001:DB8:2e50:e8::1, peer address fe80::225:4bff:fea0:6fe8
Sep 15 18:53:49 dhcp-server-1 dhcpd: Reply NA: address 2001:DB8:2e50:e8:7fff:ffff:ffff:fffe to client with duid 00:01:00:01:20:e5:6e:2d:00:25:4b:a0:6f:e8 iaid = 0 valid for 600 seconds
Sep 15 18:53:49 dhcp-server-1 dhcpd: Sending Relay-reply to 2001:DB8:2e50:e8::1 port 547
```
Using tcpdump, I can see that the client request had preferred and valid lifetimes of 0 but the reply from the server had preferred lifetime of 375 and valid lifetime of 600.
```
1505501629.303271 IP6 (class 0xe0, hlim 255, next-header UDP (17) payload length: 197) 2001:DB8:2e50:e8::1.547 > 2001:DB8:2e50:e4::226.547: [udp sum ok] dhcp6 relay-fwd (linkaddr=2001:DB8:2e50:e8::1 peeraddr=fe80::225:4bff:fea0:6fe8 (relay-message (dhcp6 request (xid=450fb (client-ID hwaddr/time type 1 time 551906861 00254ba06fe8) (option-request DNS-server DNS-search-list) (elapsed-time 0) (server-ID hwaddr/time type 1 time 542736789 00259061f77a) (IA_NA IAID:0 T1:0 T2:0 (IA_ADDR 2001:DB8:2e50:e8:7fff:ffff:ffff:fffe pltime:0 vltime:0)))) (opt_79) (interface-ID 4769302f302f312e3234...) (Remote-ID 9 0200010000f0000a0003...))
1505501629.303633 IP6 (hlim 64, next-header UDP (17) payload length: 181) 2001:DB8:2e50:e4::226.547 > 2001:DB8:2e50:e8::1.547: [udp sum ok] dhcp6 relay-reply (linkaddr=2001:DB8:2e50:e8::1 peeraddr=fe80::225:4bff:fea0:6fe8 (interface-ID 4769302f302f312e3234...) (relay-message (dhcp6 reply (xid=450fb (IA_NA IAID:0 T1:0 T2:0 (IA_ADDR 2001:DB8:2e50:e8:7fff:ffff:ffff:fffe pltime:375 vltime:600)) (client-ID hwaddr/time type 1 time 551906861 00254ba06fe8) (server-ID hwaddr/time type 1 time 542736789 00259061f77a) (DNS-server 2001:DB8:2e50:a::10 2001:DB8:2e50:a::74))))
```
It seems that option dhcp6.ia-na during the on commit {} may contain data from the client request packet instead of the server reply packet.
This seems to me like it should be considered a bug since it makes it impossible to get the lease time during on commit {}.
**Describe alternatives you've considered**
Presently I'm running tshark to extract the lease length and log it that way. This is not an ideal solution.
**Contacting you**
please contact me if you need to: perl-list at network1.netOutstandinghttps://gitlab.isc.org/isc-projects/bind9/-/issues/977Generate dlz_minimal.h from lib/dns/include/dns/clientinfo.h and others2023-11-02T16:42:09ZOndřej SurýGenerate dlz_minimal.h from lib/dns/include/dns/clientinfo.h and othersWhen keeping stuff in sync, it's very prone to break at some point in future. Instead of adding test that compares the data structures from dlz_minimal.h to their BIND library counterparts, we should rather generate dlz_minimal.h at the...When keeping stuff in sync, it's very prone to break at some point in future. Instead of adding test that compares the data structures from dlz_minimal.h to their BIND library counterparts, we should rather generate dlz_minimal.h at the build time from the pieces that needs to be included.Not plannedhttps://gitlab.isc.org/isc-projects/kea/-/issues/564Customer request: relax constraints on allowable option types to permit opti...2019-10-12T09:32:53ZMichael McNallyCustomer request: relax constraints on allowable option types to permit option type 0 (and 255?)We have a customer who would like us to relax the constraints on option types which forbid using an option with code 0, which is apparently permissible in ISC DHCP and used by one of their network equipment vendors.
When they try to co...We have a customer who would like us to relax the constraints on option types which forbid using an option with code 0, which is apparently permissible in ISC DHCP and used by one of their network equipment vendors.
When they try to configure such an option in Kea they get:
```
2019-03-28 11:23:01.204 ERROR [kea-dhcp4.dhcp4/3841] DHCP4_PARSER_FAIL failed to create
or run parser for configuration element client-classes: option data does not match option
definition (space: ZTP, code: 0): Can't create V4 option of type 0, V4 options are in
range 1..254 (/etc/kea/kea-dhcp4.conf:144:33)
```Kea1.6Tomek MrugalskiTomek Mrugalskihttps://gitlab.isc.org/isc-projects/kea/-/issues/565Client-Classes break KEA Config-Set API Call?2019-05-21T12:36:55ZGhost UserClient-Classes break KEA Config-Set API Call?---
name: Client-Classes break KEA Config-Set API Call?
about: Config-Set API call
---
**Describe the bug**
We've created a client-class for IP-phones in the KEA configuration. Now that we've started automating with the API we discover...---
name: Client-Classes break KEA Config-Set API Call?
about: Config-Set API call
---
**Describe the bug**
We've created a client-class for IP-phones in the KEA configuration. Now that we've started automating with the API we discovered an issue with the config-set API call, which reports a bad request.
The configuration is live, and KEA accepts it. However, even when we get the config through the API, and try to set it without changing it, it reports a "bad request".
**To Reproduce**
(Part of the) Config that works:
"shared-networks": [ ],
"client-classes": [
{
"name": "Innovaphone"
}
],
"subnet4": [
(Part of the) Config that doesn't work:
"shared-networks": [ ],
"client-classes": [
{
"name": "Innovaphone",
"test": "option[60].hex == '1.3.6.1.4.1.6666'",
"option-def": [
{
"name": "vendor-encapsulated-options",
"code": 43,
"type": "empty",
"encapsulate": "Innovaphone"
}
],
"option-data": [
{
"name": "h323-gatekeeper",
"code": 200,
"space": "Innovaphone",
"data": "10.90.249.1"
},
{
"name": "default-coder",
"code": 203,
"space": "Innovaphone",
"data": "G711A\\,20\\,k4/G711a\\,20\\,k4"
},
{
"name": "language",
"code": 204,
"space": "Innovaphone",
"data": "dut"
},
{
"name": "dialtone-type",
"code": 210,
"space": "Innovaphone",
"data": "0x2C"
},
{
"name": "update-URL",
"code": 215,
"space": "Innovaphone",
"data": "http://10.90.249.1/DRIVE/CF0/update/"
},
{
"name": "vendor-encapsulated-options",
"code": 43
}
]
}
],
"subnet4": [
**Expected behavior**
Being able to set the exact same config through the API that would work on the CLI.
**Environment:**
- Kea version: which release? KEA 1.4
- OS:Ubuntu 16.04 x64
**Contacting you**
gitlab / emailKea1.6Wlodzimierz WencelWlodzimierz Wencelhttps://gitlab.isc.org/isc-projects/bind9/-/issues/978Incorrect service state is reported during shutdown on Windows2019-04-19T08:19:36ZMichał KępieńIncorrect service state is reported during shutdown on WindowsWhen a Windows service receives a request to stop, it should not set its state to `SERVICE_STOPPED` until it is completely shut down as doing that allows the operating system to kill that service prematurely. In the case of `named`, thi...When a Windows service receives a request to stop, it should not set its state to `SERVICE_STOPPED` until it is completely shut down as doing that allows the operating system to kill that service prematurely. In the case of `named`, this may e.g. prevent the PID file and/or the lock file from being cleaned up.
Apparently `named` has been doing this wrong for the past 18 years.
To reproduce:
1. Install `named` as a service using `BINDInstall.exe`.
2. Put the following `named.conf` in place:
```
options {
directory "<path-to-some-writable-directory>";
pid-file "named.pid";
};
```
3. Start the "ISC BIND" service using the "Services" applet (`services.msc`).
4. Stop the "ISC BIND" service using the "Services" applet.
Chances are (it is a race) that `named.pid` will still be present in the working directory, even though it should not be. Also, the "exiting" message may not be logged.
By contrast, if `named` is stopped using `rndc stop`, shutdown always completes as intended.Michał KępieńMichał Kępieńhttps://gitlab.isc.org/isc-projects/bind9/-/issues/979named does not clean up the lock file on Windows2019-04-19T09:20:32ZMichał Kępieńnamed does not clean up the lock file on WindowsTo reproduce, prepare any working `named.conf` and then, from any writable directory, run:
```
named.exe -g -c <path-to-named.conf> -X named.lock
```
Then interrupt the process using CTRL+C. `named.lock` will still be there even thoug...To reproduce, prepare any working `named.conf` and then, from any writable directory, run:
```
named.exe -g -c <path-to-named.conf> -X named.lock
```
Then interrupt the process using CTRL+C. `named.lock` will still be there even though it should not be.Michał KępieńMichał Kępieńhttps://gitlab.isc.org/isc-projects/kea/-/issues/566DHCPv6 server to delete elements of the local configuration that were deleted...2019-04-18T16:35:39ZMarcin SiodelskiDHCPv6 server to delete elements of the local configuration that were deleted from the CBThe DHCPv6 server is able to periodically fetch new configuration elements and updated configuration elements from the configuration backend. However, it currently doesn't delete those elements of the local configuration that have been d...The DHCPv6 server is able to periodically fetch new configuration elements and updated configuration elements from the configuration backend. However, it currently doesn't delete those elements of the local configuration that have been deleted from the database. The DHCPv4 server is already doing it so it is a matter of mostly copying the logic from the v4 server.
We also have to make sure that the MySQL v6 configuration backend sets the ids of deleted elements for option definitions, options, networks and subnets.Kea1.6Thomas MarkwalderThomas Markwalderhttps://gitlab.isc.org/isc-projects/kea/-/issues/567Remove Lease::t1_ and t2_ members2019-07-22T10:51:09ZThomas MarkwalderRemove Lease::t1_ and t2_ membersCurrently our Lease class has members for t1_ and t2_, however we never actually use them other than in a few system tests. These values are not really specific to a given lease, and in fact, for v6 all leases share a common value throu...Currently our Lease class has members for t1_ and t2_, however we never actually use them other than in a few system tests. These values are not really specific to a given lease, and in fact, for v6 all leases share a common value through their IA_XX. While implementing #365 for v6 I spent a good deal of time in a rabbit hole chasing down why these members exist.
I suggest we remove them to avoid confusion and to any temptation hook developers might have to use them for something.Kea1.6-beta2Francis DupontFrancis Duponthttps://gitlab.isc.org/isc-projects/kea/-/issues/568Handle Debian/Ubuntu multiple architecture support2023-01-12T15:28:56ZFrancis DupontHandle Debian/Ubuntu multiple architecture supportCurrently in configure we look for libraries in .../lib and .../lib64 directories. It is not enough on Debian/Ubuntu which uses an architecture triplet so e.g we have to add .../lib/x86_64-linux-gnu to the directory list.
Two notes:
- ...Currently in configure we look for libraries in .../lib and .../lib64 directories. It is not enough on Debian/Ubuntu which uses an architecture triplet so e.g we have to add .../lib/x86_64-linux-gnu to the directory list.
Two notes:
- freeradius client library already uses this with the ARCH_TRIPLET variable (set to ```/x86_64-linux-gnu``` with a leading /.
- the triplet is returned by ```g++ -dumpmachine``` with only one -.
As the path can be given I am not sure there is no round around today but now the issue is identified there is no need to wait to be caught by it.backloghttps://gitlab.isc.org/isc-projects/bind9/-/issues/980util/update_copyrights now needs to handle files with CR LF endings.2019-04-12T04:28:08ZMark Andrewsutil/update_copyrights now needs to handle files with CR LF endings.https://gitlab.isc.org/isc-projects/dhcp/-/issues/7Improve error message "mdb.c(319): non-null pointer"2021-03-02T07:27:17ZCathy AlmondImprove error message "mdb.c(319): non-null pointer"Per Support ticket [RT #14122](https://support.isc.org/Ticket/Display.html?id=14122)
In a dhcpd.conf that contains both client identifier AND uid in the
same host declaration, the following warning message is emitted as dhcpd starts:
`...Per Support ticket [RT #14122](https://support.isc.org/Ticket/Display.html?id=14122)
In a dhcpd.conf that contains both client identifier AND uid in the
same host declaration, the following warning message is emitted as dhcpd starts:
`mdb.c(319): non-null pointer`
Having both is ambiguous and not supported - but the error message is not in the least helpful or useful for diagnosing what is wrong.4.4.2Thomas MarkwalderThomas Markwalderhttps://gitlab.isc.org/isc-projects/dhcp/-/issues/9DHClient: static lease not assigned with no DHCP Server responding (No DHCPOF...2019-11-18T16:38:29ZGhost UserDHClient: static lease not assigned with no DHCP Server responding (No DHCPOFFERS received.)---
name: DHC Client: Access all dynamic and static leases on signal
about: debugging DHC client
---
**Some initial questions**
- Are you sure your feature is not already implemented in the latest ISC DHCP version? YES
- Are you sure ...---
name: DHC Client: Access all dynamic and static leases on signal
about: debugging DHC client
---
**Some initial questions**
- Are you sure your feature is not already implemented in the latest ISC DHCP version? YES
- Are you sure your feature is not already implemented in the latest Kea version? YES - this is a Client issue.
- Are you sure what you would like to do is not possible using some other mechanisms?
- Have you discussed your idea on dhcp-users or dhcp-workers mailing lists? unk
**Is your feature request related to a problem? Please describe.**
Was trying to troubleshoot an issue:
In principle: dhclient does not use the static leases as of
current (and for a couple of versions) if all dynamic leases have expired
because of a programming bug in dhclient.c, routine state_panic in line 2331:
loop needs to be set to 0/NUL here, otherwise with no dynamic leases around
the variable loop is already dhclient->active when reaching line 2403 (after
jumping to activate_next), hence that static lease never gets into the while
loop and never gets activated, instead the "no leases in persistent database
- sleeping" message arrives, dhclient removes the IP and kills the network.
We currently experience such a problem where a DHCP server (by a router of an
ISP) does not respond for hours (for unknown reasons) causing havoc in the
network (which otherwise would work).
I have for now recompiled my own version with that bug fix in place which
works well now.
**Describe the solution you'd like**
I'd like to file a feature request to be able to see the internal
leases database (including dynamic and static leases with all details) upon a
signal (e.g. kill -USR1 pid of dhclient), this would be extremely useful for
such debugging purposes. So far I assumed the entry in the dhclient.conf
wasn't valid/erroneous and therefore rejected somehow.
**Describe alternatives you've considered**
**Additional context**
**Funding its development**
**Participating in development**
**Contacting you**
(request entered on behalf of the user, who had trouble setting up a valid account on Gitlab)4.4.2Thomas MarkwalderThomas Markwalder