Details
-
Type: Task
-
Status: Closed (View Workflow)
-
Priority: Major
-
Resolution: Done
-
Affects Version/s: None
-
Fix Version/s: None
-
Labels:None
-
Story Points:5
-
Epic Link:
-
Sprint:Spring 4 2023 Feb 21, Spring 9 2023 May 1, Summer 1 2023 May 15, Summer 2 2023 May 29, Summer 3 2023 June 12, Summer 4 2023 June 26, Summer 5 2023 July 10, Summer 6 2023 July 24, Summer 7 2023 Aug 7, Summer 8 2023 Aug 21
Description
Ann requested more space and fix to "igbquickload.org" resolution on-campus.
See attached email explaining details and work that was done on it by UNCC IT professional Michael Cowan:
Tasks:
- migrate data to the new location and test access
- create new git code repository for managing Integrated Genome Browser Quickload sites for the project
- create Web page with Javascript button users can click to add the new quickload site to IGB
References:
- Add drive to a VM w/o rebooting it: https://www.cyberciti.biz/tips/vmware-add-a-new-hard-disk-without-rebooting-guest.html
- Find out if your host is a VM: https://kb.vmware.com/s/article/1005436
- Add new disk to Linux (not necessarily a VM): https://www.tecmint.com/add-new-disk-to-an-existing-linux/
- Info about Red Hat distributions: https://access.redhat.com/articles/3078#RHEL9
Attachments
- email.pdf
- 175 kB
Issue Links
- relates to
-
IGBF-3395 Update readme for quickload site
- Closed
Activity
Made new repo with "main" as main branch: https://bitbucket.org/hotpollen/genome-browser-visualization/src/main/
Moving code from other repositories into this one.
Michael Cowan One IT set up the disk and configured it:
... it’s mounted at /mnt/igbdata . I would be grateful if you could log on sometime soon and verify everything looks good. I’ll delete the snapshot I made once you do.
- Disk mounting looks OK (ls, chmod, work as expected)
Next:
- Enable apache to access new disk mount
- Find out why scp installed on igbquickload.org host lacks "-J" option
- Probably need to update this VM. Not sure how. Look into it.
New error:
[root@cci-vm12 httpd]# service httpd start Starting httpd: (30)Read-only file system: httpd: could not open error log file /etc/httpd/logs/error_log. Unable to open logs [FAILED]
Asked M. Cowan for assistance.
Also, I need to find the VM host's public IP address.
Logged in to new VM. 5 Tb storage appears to be missing. Not sure if I am in the right place, or not. Emailed MC for info.
Update:
There are now two igbquickload hosts on UNCC infrastructure:
1) old one - vm12, private IP address 10.16.57.232, accessed with
ssh -J aloraine@cci-jump.uncc.edu -p 1657 aloraine@igbquickload.org
or
ssh -J aloraine@cci-jump.uncc.edu -p 1657 aloraine@10.16.57.232
2) new one - cci-igb, private IP address 10.16.80.158, accessed with
ssh aloraine@10.16.80.158
Installed and started apache2 on cci-igb. Web server is active. No configurations done as yet.
Waiting on Michael to configure the host to enable me to use the new 5 Tb of space.
Public IP external address for cci-vm12
Name: cci-vm12.uncc.edu
Address: 152.15.236.217
Name: cci-vm12.uncc.edu
Address: 10.16.57.232
Currently public IP for "igbquickload.org" is : 52.205.102.207 (an EC2 named "quickload.bioviz.org" in my account)
plan: move "mac address" associated with vm12 (old vm) over to the new one (cci-igb) to avoid having to get a new public IP address.
Notes:
Saturday morning 5:00 am updates to OS's for the VMs.
Make sure Apache is configured to restart when system reboots.
e.g., 'service apache2 enable' or similar
Suggestion from Michael Cowan for how to manage data transfers:
Try "tailscale" - a very nice tool for moving data from place to place
To use it, install client onto machines you would like to access. It functions like a vpn tunnel but with more fine-tuned management. It's a way to deal with lack of direct ssh/scp access. Uses "wireguard", a think that improves on vpn and is built into the Linux kernal.
Can use any of the nodes of an egress node.
Other notes:
Currently CCI is using "quobyt" on the slurm cluster in CCI. It is a storage solution that is flexible and can be configured by the user to accommodate whatever file system setup they need.
He suggested we could use it for igb data hosting.
Used term: "pciu4 fabric" (not sure what this is; sounds important to know about)
CCI is looking at licensing products from https://lambdalabs.com, as many faculty have been purchasing / using it. This company re-packages open source software and sells it, kind of like Red Hat, maybe?
Update:
Michael will make a change to the storage configuration and let me know when it's ready to be set up.
New storage configuration:
root@cci-igb:~# df -h Filesystem Size Used Avail Use% Mounted on tmpfs 1.2G 1.3M 1.2G 1% /run /dev/mapper/system_vg-root_lv 18G 3.3G 14G 20% / tmpfs 5.9G 0 5.9G 0% /dev/shm tmpfs 5.0M 0 5.0M 0% /run/lock /dev/sda1 1.1G 6.1M 1.1G 1% /boot/efi /dev/mapper/system_vg-home_lv 5.9G 96K 5.6G 1% /home /dev/mapper/system_vg-opt_lv 7.8G 24K 7.4G 1% /opt /dev/mapper/system_vg-var_lv 7.8G 113M 7.3G 2% /var /dev/mapper/system_vg-lib_lv 7.8G 631M 6.8G 9% /var/lib /dev/mapper/data_vg-srv_lv 2.0T 28K 1.9T 1% /srv /dev/mapper/system_vg-log_lv 7.8G 62M 7.4G 1% /var/log tmpfs 1.2G 4.0K 1.2G 1% /run/user/97303865 /dev/mapper/data_vg-igbdata_lv 4.9T 28K 4.7T 1% /mnt/igbdata
Logged into new host.
Made new virtual hosts following instructions in https://www.digitalocean.com/community/tutorials/how-to-set-up-apache-virtual-hosts-on-ubuntu-20-04.
Virtual hosts:
- data.bioviz.org
- igbquickload.org
Tested by editing my local /etc/hosts file to include:
10.16.80.158 data.bioviz.org 10.16.80.158 igbquickload.org 10.16.80.158 www.igbquickload.org
Noticed that the svn volume was unmounted on the EC2 host. Re-mounted following these notes.
Logged into the new host and use svn to checkout the IGB quickload data repository:
root@cci-igb:/mnt/igbdata# svn --username=guest co https://svn.bioviz.org/repos/genomes/quickload
Note the location is /mnt/igbdata, the location with 5 Tb of data.
contents of igbquickload (old vm) host:
[aloraine@cci-vm12 htdocs]$ ls -lh total 16K drwxr-xr-x 6 root root 4.0K Nov 11 2018 autoindex_strapdown drwxr-xr-x 3 root root 4.0K Nov 11 2018 bar lrwxrwxrwx 1 root root 14 Nov 11 2018 blueberry -> /srv/blueberry lrwxrwxrwx 1 root root 12 Nov 11 2018 chipseq -> /srv/chipseq lrwxrwxrwx 1 root root 11 Nov 11 2018 dnaseq -> /srv/dnaseq -rw-r--r-- 1 root root 357 Nov 11 2018 index.html lrwxrwxrwx 1 root root 14 Nov 11 2018 quickload -> /srv/quickload lrwxrwxrwx 1 root root 11 Nov 11 2018 rnaseq -> /srv/rnaseq lrwxrwxrwx 1 root root 16 Nov 12 2018 schallerlab -> /srv/schallerlab lrwxrwxrwx 1 root root 29 Nov 11 2018 secureQuickloadTestSites -> /srv/secureQuickloadTestSites lrwxrwxrwx 1 root root 26 Nov 11 2018 smokeTestingQuickload -> /srv/smokeTestingQuickload lrwxrwxrwx 1 root root 11 Nov 12 2018 soyseq -> /srv/soyseq drwxr-xr-x 4 root root 4.0K Sep 23 2022 styling
Moving smaller directories over to the new VM.
Copying hotpollen from RENCI host with:
root@cci-igb:/mnt/igbdata# rsync -avzhP -e "ssh -J aloraine@hop.renci.org" aloraine@lorainelab-quickload.scidas.org:/projects/igbquickload/lorainelab/www/main/htdocs/hotpollen/ hotpollen
onto:
- The new VM
- The lustre file system on the cluster
Note:
The new Quickload main directory needs a symbolic link from the web-facing root "/" to its "styling" subdirectory to enable strapdown presentation of directory contents.
Notes on today's work:
- Requested Michael Cowan to switch the network as required to make the new VM public. He will do it at 1 pm tomorrow.
- When I checked hotpollen quickload data transfer (via rsync from VM to renci) after scrum, about 1.3 T out of 1.5 T had transferred
- The process doing the transfer is taking place in a "screen" session started the night before, user root.
Notes from last night:
- When I tried to access data.bioviz.org via https, the server failed to respond. Why? Need to review logs.
Transfer complete:
sent 49.29K bytes received 1.42T bytes 24.70M bytes/sec
total size is 1.42T speedup is 1.00
Networking is updated.
Logged into the new VM with:
ssh -J aloraine@cci-jump.uncc.edu -p 1657 aloraine@10.16.57.232
Opened my browser to "10.16.57.232" and observed the index.html page set up for data.bioviz.org.
Visiting http://10.16.57.232/quickload/ or http://10.16.57.232/hotpollen opened the expected locations.
Visiting the public IP address opened nothing, no doubt because I'm on the UNC Charlotte network, a network that does not allow users within the network to access Web sites hosted by computers within the network unless their private IP addresses are used. The public IP addresses for these hosts are not recognized by computers within the network.
Next steps:
- Replicate all data hosted currently on the "igbquickload.org" EC2 host onto the new VM at UNC Charlotte.
- Once that is done, switch the IP address for "igbquickload.org" domain to the new VM's public facing IP address.
Something has changed. I now can log into the new VM with:
- ssh aloraine@10.16.57.232
No password required as my key is already installed in my account's "authorized hosts" file.
Commencing to replicate igbquickload.org content on the host.
Required directories:
- chipseq - DONE
- dnaseq - DONE
- quickload - DONE
- rnaseq - in progress
Interesting. File transfer rate is much faster. Now it's >150MB/s versus 20MB/s yesterday.
"rnaseq" is nearly finished. Decided to switch igbquickload.org to 152.15.236.217.
"rnaseq" is finished with the following report:
sent 45.73M bytes received 275.21G bytes 18.61M bytes/sec
total size is 572.68G speedup is 2.08
This is a little confusing because when I changed into the source directory on the RENCI host (/projects/igbquickload/lorainelab/www/main/htdocs/rnaseq) and executed "du -h", the total size of the current working directory (".") was reported as 655G. However, when I ran the same command in the target directory on the new VM (/mnt/igbdata/rnaseq), its size was reported as 534G. Why the different sizes?
Looking into it a bit more...
Set up SSL.
Closing as this is now done.
Copied blueberry data with:
data.bioviz.org root$ rsync -avzhP -e "ssh -J aloraine@hop.renci.org" aloraine@lorainelab-quickload.scidas.org:/projects/igbquickload/lorainelab/www/main/htdocs/blueberry/ blueberry
Copied soyseq data with:
rsync -avzhP -e "ssh -J aloraine@hop.renci.org" aloraine@lorainelab-quickload.scidas.org:/projects/igbquickload/lorainelab/www/main/htdocs/soyseq/ soyseq
Check the new disk with:
Check linux release:
Check linux distribution: