limetech Posted November 17, 2015 Share Posted November 17, 2015 Download Clicking 'Check for Updates' on the Plugins page is the preferred way to upgrade. Note: If you are using plugins, please see this thread regarding plugin support. This is another maintenance release to update a couple packages due to security advisories, add some requested kernel options, and to integrate latest work from bonienl (once again: great work and thanks!). In addition we made a small change to parity sync/check engine that should result in restoring better performance, even with mvsas cards (without using the 'nr_requests' hack - please remove this if you have it in your 'go' file). Please post here if it makes a difference for you. Sorry it took so long to get this out: we have been very busy working on 6.2-beta which I wanted to release at the same time as 6.1.4; but, we still have a couple more issues to resolve before public beta of 6.2 is ready. > unRAID Server OS Change Log Version 6.1.4 2015-11-15 </pre> <ul>curl: 7.45.0 (SSA:2015-302-01) linux: 4.1.13 linux: add kernel options (per user request): CONFIG_BE2NET: ServerEngines 10Gbps NIC - BladeEngine linux: [PATCH] ethernet: atheros driver support e2400 linux: [PATCH] scsi: mvsas support for Asus P9A-I SAS on-board controller ntp: version 4.2.8p4 (SSA:2015-302-03) ssmtp: fixed issue where '#' characters could not be used in email account passwords unraid: introduce md_sync_thresh default set to md_sync_window/2 to improve parity sync/check webGui: enable Apply button upon 'input change' per Eric webGui: fix bug in Browse.page where presence of dotfiles messed up Location column webGui: integrate numerous bergware changes: Check docker enabled state for dockerupdate process Array devices, remove unused variables PHP CLI scripts updates Monitor script update and add notifications for unassigned devices DeviceList update Improved handling of unassigned devices Correction in SMART selftest Show "please wait" message when moving thru disk pages Fixed settings selection in Mover Fixed size computation for shares with spaces in their name Save notifications "message" and make them show/hide in archive Add "Compute All" function to user and disk shares Move "Archived Notifications" to Tools Other improvements </ul> <h2>Version 6.1.3 2015-09-20</h2> <ul> firmware: version 20150820 These are firmware files for the linux kernel loaded into devices by certain drivers. There are some updates to Realtek firmware which may or may not improve certain Realtek NIC performance. linux: version 4.1.7 linux: add kernel options (all per user request): CONFIG_SENSORS_NCT7904: Nuvoton NCT7904 CONFIG_FMC_CHARDEV: FMC mezzanine driver that registers a char device CONFIG_FMC_TRIVIAL: FMC trivial mezzanine driver (software testing) CONFIG_FMC_WRITE_EEPROM: FMC mezzanine driver to write I2C EEPROM linux: removed kernel options: CONFIG_SCSI_AIC94XX: Adaptec AIC94xx SAS/SATA support This was removed because required firmware file is not redistributable. This is for an old PCI-X card anyway. If you are using this card contact support@lime-technology.com. CONFIG_EXT2_FS: Second extended fs support CONFIG_EXT3_FS: Ext3 journalling file system support Removing these options omits the code for these file systems because the ext4 file system driver now also handles ext2 and ext3 mounts. php: version 5.4.45 This is a security update. Refer to http://php.net/ChangeLog-5.php#5.4.45. webGUI changes: Remember view selection on Dashboard page Correct disk utilization on dashboard page Correction in days calculation of parity in statuscheck script Return disk PoH value as years, months, days, hours Allow SMART info for unassigned devices Corrected display of flash device in Main VM Manager: allow 128MB and 256MB memory options </ul> <h2>Version 6.1.2 2015-09-08</h2> <ul> emhttp: fix possible infinite loop if client closes HTTP connection during POST data transmission This should fix some instances of webGui hanging. emhttp: force close smb share connections if that share security mode or access list is changed Used to be that Windows would close SMB connections shortly after last I/O to a server. These days Windows likes to keep conections open. Consequently, changing a share security mode or user access list would appear to not take effect. This change fixes that behavior. linux: add kernel options: CONFIG_SCSI_MVSAS_TASKLET: Support for interrupt tasklet The was removed to test effect on mvsas slowdown issue. Since it did not fix the issue we are restoring it. slack: added a couple terminfo mappings (user request): xterm-color xterm-256color </ul> <h2>Version 6.1.1 2015-09-06</h2> <ul>emhttp: fix: for new installations: disk share export set to 'Auto', user share export set to 'Yes' docker: fix: create initial docker template repo emhttp: for new installations: disk spindown delay set to 0 (never spin down) linux: enable kernel options: CONFIG_BLK_DEV_NVME: NVM Express block device CONFIG_USB_ACM: USB Modem (CDC ACM) support linux: remove kernel options: CONFIG_SCSI_MVSAS_TASKLET: Support for interrupt tasklet linux: patch to ignore logging those useless and harmless msr errors that flood the syslog when running VMs qemu: ovmf: firmware Date: 20150904 qemu: remove unneeded UefiShell.iso file VM Manager: noVNC: performance improvements and enhancements </ul> <h2>Version 6.1.0 2015-08-31</h2> <ul>avahi: supress "Invalid response packet from host" messages btrfs-progs: version 4.1.2 docker: less verbose logging by default docker: version 1.7.1 emhttp: bridge forward delay configurable emhttp: disk share export Yes/No/Auto support emhttp: fix problem not able to start multiple btrfs scrubs emhttp: get rid of passing arbitrary commands in URL emhttp: let Pro start regardless of attached device count emhttp: let Trial support disks in any slots emhttp: properly handle non-rotational devices standby mode (ie, 'spindown') emhttp: properly handle requests from localhost (IP address 127.0.0.1) emhttp: require credentials on all accesses, not just webGui pages gnutls: 3.3.17.1 (SSA:2015-233-01) libvirt: version 1.2.18 linux: version 4.1.5 linux: support "Lite-On IT Corp. / Plextor M6e PCI Express SSD [Marvell 88SS9183] (rev 14)", see https://lkml.org/lkml/2015/2/2/226 mover: only move objects where "Use Cache" is set to "Yes" for a share ntp: enable by default ntp: 4.2.8p3 (SSA:2015-188-03) OpenELEC: added 6.0-beta3 image (beta3 includes hdmi sound fixes for nvidia cards and corrected 'addon_url' to allow plugins in Kodi once again) OpenELEC: use readonly for openelec disk image openssl: 1.0.1p (SSA:2015-190-01) php: enable pcntl php: version 5.4.44 samba: include linux "dot" objects in directory listings samba: version 4.1.19 slack: ca-certificates version 20150426 slack: check USB Flash device upon startup (set 6th field of fstab to 1) unRAIDServer.plg: improvements in unRAID-5 update: ping google instead of local gateway to determine if network up VM Manager: allow usb devices to be used that do not have a name/desciption VM Manager: Fix: OVMF+Q35 needs the bus set to 'usb' instead of 'ide' for CD-ROMs VM Manager: support OpenELEC template VM Manager: Updated noVNC (using latest as of 8/5/15, commit 8f3c0f6) webGui: Add APCUPS summary to dashboard page webGui: Add refresh button to Docker Scrub operation webGui: added "Hide dot files" control on SMB Settings page webGUI: added banner and user thumbnail image upload support webGui: additional options for text/graphical display of Used/Free. webGui: all '#command' and 'cmd' values must reference scripts within /usr/local/emhttp/ jail. webGui: Corrected: default settings were not applied for apcupsd webGui: disable Array Stop button when mover is running webGui: enhanced report generation webGui: feedback for kool kids webGui: fix issues with uploading thumbnails, banner webGui: fix docker update regression introduced in 6.1-rc1 webGui: fixed memory max display in dashboard webGui: fix notifications not initialized on reboot webGui: force browser reload of selected scripts and css files webGui: include linux "hidden" objects in directory listings webGui: install key: accept pre-keyserver links to key files webGui: integrate latest dynamix changes; many changes, refer to github webGui: Only show docker utilization when docker is enabled webGui: rearrange columns on Main page webGui: Replace alert-box for SweetAlert (enhanced messaging) webGui: use 'ftpusers' script to define ftp users webGui: use 'agent' script to control notification agents webGui: other misc. bug fixes and improvements </ul> <h2>Version 6.0.1 2015-06-26</h2> <ul>linux: enable kernel options: CONFIG_USB_PRINTER: USB Printer support CONFIG_USB_SERIAL_CP210X: USB CP210x family of UART Bridge Controllers openssl: 1.0.1o (SSA:2015-162-01) php: 5.4.41 (SSA:2015-162-02) unRAIDServer plugin (this file): be compatible with both unRAID-5 and unRAID-6 webGui: Docker: Fix: volume mappings with apostrophes not escaping correctly webGui: integrate online feedback wget: version 1.16.3 </ul> <h2>Version 6.0.0 2015-06-15</h2> <ul>docker: set default template repo webGui: other cosmetic changes </ul> <h2>Version 6.0-rc6a</h2> <ul>docker: fix bug in docker/daemon/container.go: /etc/resolv.conf permissions should be 0644 not 0600 webGui: add 'restart' to docker context menu </ul> <h2>Version 6.0-rc6</h2> <ul>dhcpcd: update to 6.8.1 dhcpcd: put dhcpcd in background if no carrier and/or no IP address after 60 seconds docker: trigger docker inotify watches on /etc/resolve.conf emhttp: fix cache devices getting unassigned when array slot count decreased slack: create informative /etc/issue file slack: maintain /etc/hostname file slack: added 'inet' command to /root - just a symlink to /etc/rc.d/rc.inet1 webGui: Info page memory display corrections; other misc. changes webGui: other misc changes (see github) </ul> <h2>Version 6.0-rc5</h2> <ul>bonding: fix to bring ethX i/f down before adding to bond; remove dhcp delay emhttp: fix disk/share status indicators emhttp: fix flash share comment not showing up, added flash share 'fsColor' emhttp: fix possible deadlock if http connection closed with output still pending emhttp: fix single-disk cache pool not mounting after format linux: enable kernel options: CONFIG_NFT_MASQ: Netfilter nf_tables masquerade support CONFIG_NFT_REDIR: Netfilter nf_tables redirect support CONFIG_NFT_MASQ_IPV4: IPv4 masquerading support for nf_tables CONFIG_NFT_REDIR_IPV4: IPv4 redirect support for nf_tables slack: fix ntp configuration handling syslinx: fix bug where Xen menu options are present in 'default' syslinux.cfg file webGui: added symlink to /usr/local/emhttp/plugins/dynamix/scripts/diagnostics in /usr/local/sbin webGui: file browswer: do not let user traverse up the tree beyond the mount point webGui: add rollover status indicator text for array status and user/disk share status webGui: add disclaimer to Diagnostics page (3rd party plugins may or may not expose personal information) webGui: custom user thumbnail/avatars support webGui: fix docker image file delete webGui: VM Manger changes (see github) </ul> <h2>Version 6.0-rc4</h2> <ul>btrfs-progs: update to 4.0.1 emhttp: fix bug where /mnt/user0 could be mounted when slot count decreased emhttp: do not count zero-size block devices against device limits emhttp: handle querystring variables with embedded single-quotes emhttp: added comment field for disk shares libvirt: patched rc.libvirt to eliminate bogus errors output to syslog libvirt: enabled virtual port capability linux: enable kernel options: CONFIG_MACVLAN: MAC-VLAN support CONFIG_MACVTAP: MAC-VLAN based tap driver CONFIG_NF_TABLES_BRIDGE: Ethernet Bridge nf_tables support CONFIG_NFT_BRIDGE_META: Netfilter nf_table bridge meta support CONFIG_BRIDGE_NF_EBTABLES: Ethernet Bridge tables (ebtables) support CONFIG_BRIDGE_EBT_*: (19 total Ethernet bridge tables) qemu: upgraded to SeaBIOS 1.8.1 shfs: add NOCOW share support unRAIDServer plugin: preserve only "previous" release; removing current release will restore previous release webGui: fix Active Directory export settings not being active webGui: let user configure share COW flag webGui: make apcupsd driver_loaded script executable webGui: VM Manager: added Force Stop context menu option when VM is paused webGui: VM Manager: Fix: raw disk size display and adjustment capabilities webGUI: Docker: Fix: correctly calculate total data pulled stat webGUI: Docker: permit image to exist on user share webGUI: create Disk Shares page under Shares, move disk share settings webGUI: correct handling of sinqle quotes in share names webGUI: other misc changes/bug fixes </ul> <h2>Version 6.0-rc3</h2> <ul>emhttp: fix Maintenance-mode array Start emhttp: properly output 'progress' strings output by emhttp_event scripts emhttp: let Pro start regardless of attached device count linux: update to 4.0.4 (btrfs fix from 4.0.3, spinlock fix from 4.0.4, btrfs balance patch) webGui: VM Manager bug fixes webGui: integrated improvements to apcupsd component </ul> <h2>Version 6.0-rc2</h2> <ul>emhttp: restore default number of array slots to 24 (from 5 introduced in -rc1) slack: add package: patch version 2.7.4 webGui: expunge all Xen "hooks" webGui: VM Manager bug fixes (refer to commit history on github) </ul> <h2>Version 6.0-rc1</h2> <ul>docker: update to 1.6.2 webGui: support replacement keys webGui: must be Stopped to change Identifcation settings </ul Quote Link to comment
sparklyballs Posted November 17, 2015 Share Posted November 17, 2015 Updated and rebooted, with nothing untoward to report so far. little surprised there hasn't been an update to the docker version in this release though. Quote Link to comment
interwebtech Posted November 17, 2015 Share Posted November 17, 2015 Updated via plug-in page. Sorta long pause before showing the login prompt on console, otherwise no issues to report. Quote Link to comment
gubbgnutten Posted November 17, 2015 Share Posted November 17, 2015 Cool. The plugin update method actually worked properly for me for the first time in a couple of releases. Looking forward to see how the parity check speed turns out now, and I'm sure the 6.2 beta will be yummy. Quote Link to comment
johnodon Posted November 17, 2015 Share Posted November 17, 2015 Upgraded without issue. Quote Link to comment
JorgeB Posted November 18, 2015 Share Posted November 18, 2015 Did a few quick parity check tests with 8 x SSDs, looks promising so far: SASLP: 71MB/s to 80MB/s (back to V5 speeds) SAS2LP: 92MB/s to 256MB/s Few other of things I noticed: nr_requests has no effect for me with v6.1.4 optimal tunable settings are very different from v6.1.3, everyone with a SASLP or SAS2LP should re-run tunables tester besides these 2 controllers also tested one Dell H310, speed decreased from 326MB/s to 248MB/s, I’m not complaining, 200MB/s+ is enough for any HDD, just in case it interests LT SAS2LP was faster with v6.1.3 and nr_requests=8 (307MB/s vs 256Mb/s), again not complaining, just reporting Will do some more tests during the weekend. Quote Link to comment
Darqfallen Posted November 18, 2015 Share Posted November 18, 2015 Does this release also include the update to the openelec image to release? Quote Link to comment
garycase Posted November 18, 2015 Share Posted November 18, 2015 Did this update by any chance fix the "Trust Parity" option so a parity check isn't started when you start the array? Quote Link to comment
limetech Posted November 18, 2015 Author Share Posted November 18, 2015 Did a few quick parity check tests with 8 x SSDs, looks promising so far: SASLP: 71MB/s to 80MB/s (back to V5 speeds) SAS2LP: 92MB/s to 256MB/s Few other of things I noticed: nr_requests has no effect for me with v6.1.4 optimal tunable settings are very different from v6.1.3, everyone with a SASLP or SAS2LP should re-run tunables tester besides these 2 controllers also tested one Dell H310, speed decreased from 326MB/s to 248MB/s, I’m not complaining, 200MB/s+ is enough for any HDD, just in case it interests LT SAS2LP was faster with v6.1.3 and nr_requests=8 (307MB/s vs 256Mb/s), again not complaining, just reporting Will do some more tests during the weekend. That's a great report, thank you! There is a tunable called 'md_sync_thresh' which you can set with a command such as: mdcmd set md_sync_thresh <N> The value <N> is a count of 'stripes' from 0..md_sync_window. The way it works is: when a parity sync/check starts, it queues up 'md_sync_window' number of 'stripes'. Thus the maximum number of in-progress 'stripes' is md_sync_window. As these 'stripes' complete the number in-progress obviously starts decreasing. When it reaches 'md_sync_thresh' a thread awakens, queuing up more 'stripes' until md_sync_window number of 'stripes' are again in-progress. The default value for 'md_sync_thresh' is md_sync_window/2. In pre-6.1.4 it behaved as if md_sync_thresh was equal to md_sync_window-1 (that is immediately queued new 'stripe' each time in-progress 'stripe' completed). In upcoming 6.2 this setting is configurable on Disk Settings page. Quote Link to comment
TODDLT Posted November 18, 2015 Share Posted November 18, 2015 Thank you Tom!! Quick report: (just ran a parity check to around 75 GB). 6.1.2 to 6.1.4 SASLP went from 70's to over 105(ish). SAS2LP went from in the 50's to currently 130. Extremely pleased. I've never had speeds that high. Whatever you tweaked, awesome job. I'll let the SAS2LP run it's course tonight and see where it winds up. Thanks again!! (hey tomorrow I'll go mark my SAS2LP thread solved) Quote Link to comment
hackztor Posted November 18, 2015 Share Posted November 18, 2015 Parity check going nice with SAS2LP. 6.1.3 60 MB/s 6.1.4 110 MB/s Quote Link to comment
JorgeB Posted November 18, 2015 Share Posted November 18, 2015 That's a great report, thank you! There is a tunable called 'md_sync_thresh' which you can set with a command such as: mdcmd set md_sync_thresh <N> The value <N> is a count of 'stripes' from 0..md_sync_window. The way it works is: when a parity sync/check starts, it queues up 'md_sync_window' number of 'stripes'. Thus the maximum number of in-progress 'stripes' is md_sync_window. As these 'stripes' complete the number in-progress obviously starts decreasing. When it reaches 'md_sync_thresh' a thread awakens, queuing up more 'stripes' until md_sync_window number of 'stripes' are again in-progress. The default value for 'md_sync_thresh' is md_sync_window/2. In pre-6.1.4 it behaved as if md_sync_thresh was equal to md_sync_window-1 (that is immediately queued new 'stripe' each time in-progress 'stripe' completed). In upcoming 6.2 this setting is configurable on Disk Settings page. Me not changing that tunable was the cause of the high-end loss of performance, good you going to make it tunable as different controllers like very different values, changing it brings maximum performance back to what I believe is close to the bandwidth limit of these controllers with 8 drives: Dell H310 likes the old value, md_sync_window-1 V6.1.3 – 326MB/s V6.1.4 – 321MB/s SAS2LP likes the new value, md_sync_window/2 V6.1.3 with nr_reqs=8 – 307MB/s V6.1.4 – 305MB/s With this and the other reports on this thread I believe the issue is fixed, great work! Now bring V6.2 Quote Link to comment
billington.mark Posted November 18, 2015 Share Posted November 18, 2015 Does this release also include the update to the openelec image to release? No, this must be in 6.2. hopefully its not too far away as im also awaiting some fixes to improve SSD performance in VMs. Quote Link to comment
dikkiedirk Posted November 18, 2015 Share Posted November 18, 2015 That's a great report, thank you! There is a tunable called 'md_sync_thresh' which you can set with a command such as: mdcmd set md_sync_thresh <N> The value <N> is a count of 'stripes' from 0..md_sync_window. The way it works is: when a parity sync/check starts, it queues up 'md_sync_window' number of 'stripes'. Thus the maximum number of in-progress 'stripes' is md_sync_window. As these 'stripes' complete the number in-progress obviously starts decreasing. When it reaches 'md_sync_thresh' a thread awakens, queuing up more 'stripes' until md_sync_window number of 'stripes' are again in-progress. The default value for 'md_sync_thresh' is md_sync_window/2. In pre-6.1.4 it behaved as if md_sync_thresh was equal to md_sync_window-1 (that is immediately queued new 'stripe' each time in-progress 'stripe' completed). In upcoming 6.2 this setting is configurable on Disk Settings page. Me not changing that tunable was the cause of the high-end loss of performance, good you going to make it tunable as different controllers like very different values, changing it brings maximum performance back to what I believe is close to the bandwidth limit of these controllers with 8 drives: Dell H310 likes the old value, md_sync_window-1 V6.1.3 – 326MB/s V6.1.4 – 321MB/s SAS2LP likes the new value, md_sync_window/2 V6.1.3 with nr_reqs=8 – 307MB/s V6.1.4 – 305MB/s With this and the other reports on this thread I believe the issue is fixed, great work! Now bring V6.2 I updated my mainserver to 6.1.4. Saw much lower parity check speeds immediately after reboot. The M1015s in my server after comparable to the H310 I think. Should I enter mdcmd set md_sync_thresh -1 from the command-line? Quote Link to comment
Ockingshay Posted November 18, 2015 Share Posted November 18, 2015 parity is definitely faster, and i would like to try and improve it some more. On V5, i used this a long time ago. Is this still the tool to use to find better tunable attributes? http://lime-technology.com/forum/index.php?topic=29009.0 Quote Link to comment
JorgeB Posted November 18, 2015 Share Posted November 18, 2015 I updated my mainserver to 6.1.4. Saw much lower parity check speeds immediately after reboot. The M1015s in my server after comparable to the H310 I think. Should I enter mdcmd set md_sync_thresh -1 from the command-line? M1015 should be the same as H310. Go to > settings > disk settings and check you md_sync_window value, default is 384. Use that value-1 for md_sync_thresh, for the default value would be: mdcmd set md_sync_thresh 383 Quote Link to comment
dikkiedirk Posted November 18, 2015 Share Posted November 18, 2015 Thanks for clearing that up! I miss-interpreted it and stupidly just set it to -1! Perhaps that is why I get lower speeds than before. Quote Link to comment
dikkiedirk Posted November 18, 2015 Share Posted November 18, 2015 Thanks for clearing that up! I miss-interpreted it and stupidly just set it to -1! Perhaps that is why I get lower speeds than before. And back to previous speeds after setting the tresh value correctly. Parity check speeds is held back by the slowest drive in the array. I don't thinks the tunables will help to further improve it. It gave a md sync window of 2480 earlier when it was on 6.1.0 Quote Link to comment
binhex Posted November 18, 2015 Share Posted November 18, 2015 I updated my mainserver to 6.1.4. Saw much lower parity check speeds immediately after reboot. The M1015s in my server after comparable to the H310 I think. Should I enter mdcmd set md_sync_thresh -1 from the command-line? M1015 should be the same as H310. Go to > settings > disk settings and check you md_sync_window value, default is 384. Use that value-1 for md_sync_thresh, for the default value would be: mdcmd set md_sync_thresh 383 im assuming the value for this setting is written to the flash drive yes?, as in it will be persistent and doesn't require re-applying on boot? Quote Link to comment
wgstarks Posted November 18, 2015 Share Posted November 18, 2015 parity is definitely faster, and i would like to try and improve it some more. On V5, i used this a long time ago. Is this still the tool to use to find better tunable attributes? http://lime-technology.com/forum/index.php?topic=29009.0 Yes. Unraid tunables tester can help you tweak at little more speed out of v6. Quote Link to comment
JorgeB Posted November 18, 2015 Share Posted November 18, 2015 im assuming the value for this setting is written to the flash drive yes?, as in it will be persistent and doesn't require re-applying on boot? I believe the setting md_sync_thresh will not survive a reboot and when unraid starts it will always be 192, half the default md_sync_window value of 384. Note that although some controllers like the Dell H310 appear to work faster with the old setting, the improvement I felt was with SSDs only, the new default settings should be ok for normal HDDs up to 250MB/s, also Tom already said this will be a tunable setting in V6.2. For the moment and if you need to change it you could add it to your go file. Quote Link to comment
JorgeB Posted November 18, 2015 Share Posted November 18, 2015 parity is definitely faster, and i would like to try and improve it some more. On V5, i used this a long time ago. Is this still the tool to use to find better tunable attributes? http://lime-technology.com/forum/index.php?topic=29009.0 I found that running tunables-tester makes much more difference on Marvell controllers, especially the SAS2LP where it can more than double the speed, naturally there’s no harm in trying but at least for me there’s zero to little difference using this on for example the Dell H310. For the SAS2LP try it with the default md_sync_thresh value of 192 and a higher value, e.g. 1024, as it should improve speeds with md_sync_window values above 1024. Quote Link to comment
Vocatus Posted November 18, 2015 Share Posted November 18, 2015 Does a point release upgrade like this require a server reboot? Thanks for the great work, seriously Quote Link to comment
squirrellydw Posted November 18, 2015 Share Posted November 18, 2015 Working great here. Thanks and keep up the great work Quote Link to comment
trurl Posted November 18, 2015 Share Posted November 18, 2015 Does a point release upgrade like this require a server reboot? Thanks for the great work, seriously Yes. The update just copies it to your flash. It won't be loaded until reboot. Quote Link to comment
Recommended Posts
Join the conversation
You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.